https://github.com/zhanghm1995/Forge_VFM4AD
A comprehensive survey of forging vision foundation models for autonomous driving, including challenges, methodologies, and opportunities.
https://github.com/zhanghm1995/Forge_VFM4AD
3dgs adaptation autonomous-driving diffusion end-to-end-autonomous-driving foundation-model large-language-models nerf pre-training survey world-models
Last synced: 11 months ago
JSON representation
A comprehensive survey of forging vision foundation models for autonomous driving, including challenges, methodologies, and opportunities.
- Host: GitHub
- URL: https://github.com/zhanghm1995/Forge_VFM4AD
- Owner: zhanghm1995
- Created: 2023-11-25T03:00:40.000Z (over 2 years ago)
- Default Branch: master
- Last Pushed: 2024-07-01T02:08:45.000Z (almost 2 years ago)
- Last Synced: 2024-07-05T14:35:12.149Z (almost 2 years ago)
- Topics: 3dgs, adaptation, autonomous-driving, diffusion, end-to-end-autonomous-driving, foundation-model, large-language-models, nerf, pre-training, survey, world-models
- Homepage:
- Size: 34.8 MB
- Stars: 206
- Watchers: 7
- Forks: 7
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
- Awesome-Segment-Anything - [code
README
# Forging Vision Foundation Models for Autonomous Driving: Challenges, Methodologies, and Opportunities
[](https://github.com/sindresorhus/awesome) [](https://github.com/Thinklab-SJTU/Awesome-LLM4AD/blob/main/LICENSE)
This is the partner repository for the survey paper [Forging Vision Foundation Models for Autonomous Driving: Challenges, Methodologies, and Opportunities](https://arxiv.org/abs/2401.08045). The repository will be continuously updated to track the progress of forging VFMs for AD. We hope this repository can act as a quick reference for researchers who wish to read the relevant papers and implement the associated methods.
**Authors**: [Xu Yan](https://yanx27.github.io/), [Haiming Zhang](https://scholar.google.com/citations?user=RPWMpyIAAAAJ&hl=zh-CN), [Yingjie Cai](https://yjcaimeow.github.io/), Jingming Guo, Weichao Qiu, Bin
Gao, Kaiqiang Zhou, Yue Zhao, Huan Jin, Jiantao Gao, [Zhen Li](https://mypage.cuhk.edu.cn/academics/lizhen/), Lihui
Jiang, Wei Zhang, Hongbo Zhang, [Dengxin Dai](https://vas.mpi-inf.mpg.de/dengxin/) and Bingbing Liu.
Our survey at a glance.
Research tree of forging vision foundation models for autonomous driving.
**NOTE**: Here we have select a number of featured papers for each part, and almost for each paper we have included the abstract and a figure from the original paper, showing the main framework or motivations, to help us take a glance about these papers (You can expand the **Abstract** button to see them). **More papers list and details can be found in our survey paper**.
⭐ **Welcome to star this repository!** ⭐
We greatly appreciate any contributions via PRs, issues, emails, or other methods.
## Citation
If this work is helpful for your research, please consider citing the following BibTeX entry :
``` bibtex
@misc{yan2024forging,
title={Forging Vision Foundation Models for Autonomous Driving: Challenges, Methodologies, and Opportunities},
author={Xu Yan and Haiming Zhang and Yingjie Cai and Jingming Guo and Weichao Qiu and Bin Gao and Kaiqiang Zhou and Yue Zhao and Huan Jin and Jiantao Gao and Zhen Li and Lihui Jiang and Wei Zhang and Hongbo Zhang and Dengxin Dai and Bingbing Liu},
year={2024},
eprint={2401.08045},
archivePrefix={arXiv},
primaryClass={cs.CV}
}
```
## 📰 News
- `[2024/07/01]` Add two 3DGS papers.
- `[2024/02/18]` Add more related papers, including a new related survey paper.
- `[2024/01/18]` Add two new data preparation related papers.
- `[2024/01/17]` Release this repository and open-access our survey paper in arXiv.
- `[2023/12/19]` Initial commit.
## Table of Content
- [Related Survey Papers](#related-survey-papers)
- [Data Preparation](#data-preparation)
- [GAN](#gan)
- [Diffusion](#diffusion)
- [NeRF](#nerf)
- [3D Gaussian Splatting](#3d-gaussian-splatting)
- [Self-supervised Training](#self-supervised-training)
- [Contrastive](#contrastive)
- [Reconstruction](#reconstruction)
- [Distillation](#distillation)
- [Rendering](#rendering)
- [World Model](#world-model)
- [Adaptation](#adaptation)
- [Vision Foundation Models](#vision-foundation-models)
- [Large Language Models](#large-language-models)
- [Multimodal Foundation Models](#multimodal-foundation-models)
## Related Survey Papers
- **A Survey for Foundation Models in Autonomous Driving**.
Abstract
The advent of foundation models has revolutionized the fields of natural language processing and computer vision, paving the way for their application in autonomous driving (AD). This survey presents a comprehensive review of more than 40 research papers, demonstrating the role of foundation models in enhancing AD. Large language models contribute to planning and simulation in AD, particularly through their proficiency in reasoning, code generation and translation. In parallel, vision foundation models are increasingly adapted for critical tasks such as 3D object detection and tracking, as well as creating realistic driving scenarios for simulation and testing. Multi-modal foundation models, integrating diverse inputs, exhibit exceptional visual understanding and spatial reasoning, crucial for end-to-end AD. This survey not only provides a structured taxonomy, categorizing foundation models based on their modalities and functionalities within the AD domain but also delves into the methods employed in current research. It identifies the gaps between existing foundation models and cutting-edge AD approaches, thereby charting future research directions and proposing a roadmap for bridging these gaps.
[](https://arxiv.org/abs/2402.01105)
- **Foundation Models in Robotics: Applications, Challenges, and the Future**.
Abstract
We survey applications of pretrained foundation models in robotics. Traditional deep learning models in robotics are trained on small datasets tailored for specific tasks, which limits their adaptability across diverse applications. In contrast, foundation models pretrained on internet-scale data appear to have superior generalization capabilities, and in some instances display an emergent ability to find zero-shot solutions to problems that are not present in the training data. Foundation models may hold the potential to enhance various components of the robot autonomy stack, from perception to decision-making and control. For example, large language models can generate code or provide common sense reasoning, while vision-language models enable open-vocabulary visual recognition. However, significant open research challenges remain, particularly around the scarcity of robot-relevant training data, safety guarantees and uncertainty quantification, and real-time execution. In this survey, we study recent papers that have used or built foundation models to solve robotics problems. We explore how foundation models contribute to improving robot capabilities in the domains of perception, decision-making, and control. We discuss the challenges hindering the adoption of foundation models in robot autonomy and provide opportunities and potential pathways for future advancements.
[](https://arxiv.org/abs/2312.07843) [](https://github.com/robotics-survey/Awesome-Robotics-Foundation-Models)
- **Applications of Large Scale Foundation Models for Autonomous Driving**.
Abstract
Since DARPA Grand Challenges (rural) in 2004/05 and Urban Challenges in 2007, autonomous driving has been the most active field of AI applications. Recently powered by large language models (LLMs), chat systems, such as chatGPT and PaLM, emerge and rapidly become a promising direction to achieve artificial general intelligence (AGI) in natural language processing (NLP). There comes a natural thinking that we could employ these abilities to reformulate autonomous driving. By combining LLM with foundation models, it is possible to utilize the human knowledge, commonsense and reasoning to rebuild autonomous driving systems from the current long-tailed AI dilemma. In this paper, we investigate the techniques of foundation models and LLMs applied for autonomous driving, categorized as simulation, world model, data annotation and planning or E2E solutions etc.
[](https://arxiv.org/abs/2311.12144)
- **Vision Language Models in Autonomous Driving and Intelligent Transportation Systems**.
Abstract
TODOThe applications of Vision-Language Models (VLMs) in the fields of Autonomous Driving (AD) and Intelligent Transportation Systems (ITS) have attracted widespread attention due to their outstanding performance and the ability to leverage Large Language Models (LLMs). By integrating language data, the vehicles, and transportation systems are able to deeply understand real-world environments, improving driving safety and efficiency. In this work, we present a comprehensive survey of the advances in language models in this domain, encompassing current models and datasets. Additionally, we explore the potential applications and emerging research directions. Finally, we thoroughly discuss the challenges and research gap. The paper aims to provide researchers with the current work and future trends of VLMs in AD and ITS.
[](https://arxiv.org/abs/2310.14414)
- **A Comprehensive Survey on Segment Anything Model for Vision and Beyond**.
Abstract
Artificial intelligence (AI) is evolving towards artificial general intelligence, which refers to the ability of an AI system to perform a wide range of tasks and exhibit a level of intelligence similar to that of a human being. This is in contrast to narrow or specialized AI, which is designed to perform specific tasks with a high degree of efficiency. Therefore, it is urgent to design a general class of models, which we term foundation models, trained on broad data that can be adapted to various downstream tasks. The recently proposed segment anything model (SAM) has made significant progress in breaking the boundaries of segmentation, greatly promoting the development of foundation models for computer vision. To fully comprehend SAM, we conduct a survey study. As the first to comprehensively review the progress of segmenting anything task for vision and beyond based on the foundation model of SAM, this work focuses on its applications to various tasks and data types by discussing its historical development, recent progress, and profound impact on broad applications. We first introduce the background and terminology for foundation models including SAM, as well as state-of-the-art methods contemporaneous with SAM that are significant for segmenting anything task. Then, we analyze and summarize the advantages and limitations of SAM across various image processing applications, including software scenes, real-world scenes, and complex scenes. Importantly, many insights are drawn to guide future research to develop more versatile foundation models and improve the architecture of SAM. We also summarize massive other amazing applications of SAM in vision and beyond.
[](https://arxiv.org/abs/2305.08196) [](https://github.com/liliu-avril/Awesome-Segment-Anything)
- **Foundation Models for Decision Making: Problems, Methods, and Opportunities**.
Abstract
Foundation models pretrained on diverse data at scale have demonstrated extraordinary capabilities in a wide range of vision and language tasks. When such models are deployed in real world environments, they inevitably interface with other entities and agents. For example, language models are often used to interact with human beings through dialogue, and visual perception models are used to autonomously navigate neighborhood streets. In response to these developments, new paradigms are emerging for training foundation models to interact with other agents and perform long-term reasoning. These paradigms leverage the existence of ever-larger datasets curated for multimodal, multitask, and generalist interaction. Research at the intersection of foundation models and decision making holds tremendous promise for creating powerful new systems that can interact effectively across a diverse range of applications such as dialogue, autonomous driving, healthcare, education, and robotics. In this manuscript, we examine the scope of foundation models for decision making, and provide conceptual tools and technical background for understanding the problem space and exploring new research directions. We review recent approaches that ground foundation models in practical decision making applications through a variety of methods such as prompting, conditional generative modeling, planning, optimal control, and reinforcement learning, and discuss common challenges and open problems in the field.
[](https://arxiv.org/abs/2303.04129)
- **A Survey of Large Language Models for Autonomous Driving**.
Abstract
Autonomous driving technology, a catalyst for revolutionizing transportation and urban mobility, has the tend to transition from rule-based systems to data-driven strategies. Traditional module-based systems are constrained by cumulative errors among cascaded modules and inflexible pre-set rules. In contrast, end-to-end autonomous driving systems have the potential to avoid error accumulation due to their fully data-driven training process, although they often lack transparency due to their ``black box" nature, complicating the validation and traceability of decisions. Recently, large language models (LLMs) have demonstrated abilities including understanding context, logical reasoning, and generating answers. A natural thought is to utilize these abilities to empower autonomous driving. By combining LLM with foundation vision models, it could open the door to open-world understanding, reasoning, and few-shot learning, which current autonomous driving systems are lacking. In this paper, we systematically review a research line about \textit{Large Language Models for Autonomous Driving (LLM4AD)}. This study evaluates the current state of technological advancements, distinctly outlining the principal challenges and prospective directions for the field.
[](https://arxiv.org/abs/2304.10406) [](https://github.com/Thinklab-SJTU/Awesome-LLM4AD)
- **On the Opportunities and Risks of Foundation Models**.
Abstract
AI is undergoing a paradigm shift with the rise of models (e.g., BERT, DALL-E, GPT-3) that are trained on broad data at scale and are adaptable to a wide range of downstream tasks. We call these models foundation models to underscore their critically central yet incomplete character. This report provides a thorough account of the opportunities and risks of foundation models, ranging from their capabilities (e.g., language, vision, robotics, reasoning, human interaction) and technical principles(e.g., model architectures, training procedures, data, systems, security, evaluation, theory) to their applications (e.g., law, healthcare, education) and societal impact (e.g., inequity, misuse, economic and environmental impact, legal and ethical considerations). Though foundation models are based on standard deep learning and transfer learning, their scale results in new emergent capabilities,and their effectiveness across so many tasks incentivizes homogenization. Homogenization provides powerful leverage but demands caution, as the defects of the foundation model are inherited by all the adapted models downstream. Despite the impending widespread deployment of foundation models, we currently lack a clear understanding of how they work, when they fail, and what they are even capable of due to their emergent properties. To tackle these questions, we believe much of the critical research on foundation models will require deep interdisciplinary collaboration commensurate with their fundamentally sociotechnical nature.
[](https://arxiv.org/abs/2108.07258) [](https://crfm.stanford.edu/report.html)
👆 [Back to Top](#Table-of-Content)
## Data Preparation
### GAN
- **DriveGAN: Towards a Controllable High-Quality Neural Simulation**.
Abstract
Realistic simulators are critical for training and verifying robotics systems. While most of the contemporary simulators are hand-crafted, a scaleable way to build simulators is to use machine learning to learn how the environment behaves in response to an action, directly from data. In this work, we aim to learn to simulate a dynamic environment directly in pixel-space, by watching unannotated sequences of frames and their associated action pairs. We introduce a novel high-quality neural simulator referred to as DriveGAN that achieves controllability by disentangling different components without supervision. In addition to steering controls, it also includes controls for sampling features of a scene, such as the weather as well as the location of non-player objects. Since DriveGAN is a fully differentiable simulator, it further allows for re-simulation of a given video sequence, offering an agent to drive through a recorded scene again, possibly taking different actions. We train DriveGAN on multiple datasets, including 160 hours of real-world driving data. We showcase that our approach greatly surpasses the performance of previous data-driven simulators, and allows for new features not explored before.
[](https://arxiv.org/abs/2104.15060) [](https://research.nvidia.com/labs/toronto-ai/DriveGAN/)
- **SurfelGAN: Synthesizing Realistic Sensor Data for Autonomous Driving**.
Abstract
Autonomous driving system development is critically dependent on the ability to replay complex and diverse traffic scenarios in simulation. In such scenarios, the ability to accurately simulate the vehicle sensors such as cameras, lidar or radar is essential. However, current sensor simulators leverage gaming engines such as Unreal or Unity, requiring manual creation of environments, objects and material properties. Such approaches have limited scalability and fail to produce realistic approximations of camera, lidar, and radar data without significant additional work. In this paper, we present a simple yet effective approach to generate realistic scenario sensor data, based only on a limited amount of lidar and camera data collected by an autonomous vehicle. Our approach uses texture-mapped surfels to efficiently reconstruct the scene from an initial vehicle pass or set of passes, preserving rich information about object 3D geometry and appearance, as well as the scene conditions. We then leverage a SurfelGAN network to reconstruct realistic camera images for novel positions and orientations of the self-driving vehicle and moving objects in the scene. We demonstrate our approach on the Waymo Open Dataset and show that it can synthesize realistic camera data for simulated scenarios. We also create a novel dataset that contains cases in which two self-driving vehicles observe the same scene at the same time. We use this dataset to provide additional evaluation and demonstrate the usefulness of our SurfelGAN model.
[](https://arxiv.org/abs/2005.03844)
- **GAN-Based LiDAR Translation between Sunny and Adverse Weather for Autonomous Driving and Driving Simulation**.
Abstract
Autonomous driving requires robust and highly accurate perception technologies. Various deep learning algorithms based on only image processing satisfy this requirement, but few such algorithms are based on LiDAR. However, images are only one part of the perceptible sensors in an autonomous driving vehicle; LiDAR is also essential for the recognition of driving environments. The main reason why there exist few deep learning algorithms based on LiDAR is a lack of data. Recent translation technology using generative adversarial networks (GANs) has been proposed to deal with this problem. However, these technologies focus on only image-to-image translation, although a lack of data occurs more often with LiDAR than with images. LiDAR translation technology is required not only for data augmentation, but also for driving simulation, which allows algorithms to practice driving as if they were commanding a real vehicle, before doing so in the real world. In other words, driving simulation is a key technology for evaluating and verifying algorithms which are practically applied to vehicles. In this paper, we propose a GAN-based LiDAR translation algorithm for autonomous driving and driving simulation. It is the first LiDAR translation approach that can deal with various types of weather that are based on an empirical approach. We tested the proposed method on the JARI data set, which was collected under various adverse weather scenarios with diverse precipitation and visible distance settings. The proposed method was also applied to the real-world Spain data set. Our experimental results demonstrate that the proposed method can generate realistic LiDAR data under adverse weather conditions.
[](https://www.mdpi.com/1424-8220/22/14/5287)
- **LiDAR Sensor modeling and Data augmentation with GANs for Autonomous driving**.
Abstract
In the autonomous driving domain, data collection and annotation from real vehicles are expensive and sometimes unsafe. Simulators are often used for data augmentation, which requires realistic sensor models that are hard to formulate and model in closed forms. Instead, sensors models can be learned from real data. The main challenge is the absence of paired data set, which makes traditional supervised learning techniques not suitable. In this work, we formulate the problem as image translation from unpaired data and employ CycleGANs to solve the sensor modeling problem for LiDAR, to produce realistic LiDAR from simulated LiDAR (sim2real). Further, we generate high-resolution, realistic LiDAR from lower resolution one (real2real). The LiDAR 3D point cloud is processed in Bird-eye View and Polar 2D representations. The experimental results show a high potential of the proposed approach.
[](https://arxiv.org/abs/1905.07290) [](https://github.com/ahmadelsallab/lidargan)
- **DeepRoad: GAN-based Metamorphic Autonomous Driving System Testing**.
Abstract
While Deep Neural Networks (DNNs) have established the fundamentals of DNN-based autonomous driving systems, they may exhibit erroneous behaviors and cause fatal accidents. To resolve the safety issues of autonomous driving systems, a recent set of testing techniques have been designed to automatically generate test cases, e.g., new input images transformed from the original ones. Unfortunately, many such generated input images often render inferior authenticity, lacking accurate semantic information of the driving scenes and hence compromising the resulting efficacy and reliability.In this paper, we propose DeepRoad, an unsupervised framework to automatically generate large amounts of accurate driving scenes to test the consistency of DNN-based autonomous driving systems across different scenes. In particular, DeepRoad delivers driving scenes with various weather conditions (including those with rather extreme conditions) by applying the Generative Adversarial Networks (GANs) along with the corresponding real-world weather scenes. Moreover, we have implemented DeepRoad to test three well-recognized DNN-based autonomous driving systems. Experimental results demonstrate that DeepRoad can detect thousands of behavioral inconsistencies in these systems.
[](https://arxiv.org/abs/1802.02295)
- **Deep generative modeling of lidar data**.
Abstract
Building models capable of generating structured output is a key challenge for AI and robotics. While generative models have been explored on many types of data, little work has been done on synthesizing lidar scans, which play a key role in robot mapping and localization. In this work, we show that one can adapt deep generative models for this task by unravelling lidar scans into a 2D point map. Our approach can generate high quality samples, while simultaneously learning a meaningful latent representation of the data. We demonstrate significant improvements against state-of-the-art point cloud generation methods. Furthermore, we propose a novel data representation that augments the 2D signal with absolute positional information. We show that this helps robustness to noisy and imputed input; the learned model can recover the underlying lidar scan from seemingly uninformative data.
[](https://arxiv.org/abs/1812.01180) [](https://github.com/pclucas14/lidar_generation)
👆 [Back to Top](#Table-of-Content)
### Diffusion
- **WoVoGen: World Volume-aware Diffusion for Controllable Multi-camera Driving Scene Generation**.
Abstract
Generating multi-camera street-view videos is critical for augmenting autonomous driving datasets, addressing the urgent demand for extensive and varied data. Due to the limitations in diversity and challenges in handling lighting conditions, traditional rendering-based methods are increasingly being supplanted by diffusion-based methods. However, a significant challenge in diffusion-based methods is ensuring that the generated sensor data preserve both intra-world consistency and inter-sensor coherence. To address these challenges, we combine an additional explicit world volume and propose the World Volume-aware Multi-camera Driving Scene Generator (WoVoGen). This system is specifically designed to leverage 4D world volume as a foundational element for video generation. Our model operates in two distinct phases: (i) envisioning the future 4D temporal world volume based on vehicle control sequences, and (ii) generating multi-camera videos, informed by this envisioned 4D temporal world volume and sensor interconnectivity. The incorporation of the 4D world volume empowers WoVoGen not only to generate high-quality street-view videos in response to vehicle control inputs but also to facilitate scene editing tasks.
[](https://arxiv.org/abs/2312.02934) [](https://github.com/fudan-zvg/WoVoGen)
- **Panacea: Panoramic and Controllable Video Generation for Autonomous Driving**.
Abstract
The field of autonomous driving increasingly demands high-quality annotated training data. In this paper, we propose Panacea, an innovative approach to generate panoramic and controllable videos in driving scenarios, capable of yielding an unlimited numbers of diverse, annotated samples pivotal for autonomous driving advancements. Panacea addresses two critical challenges: 'Consistency' and 'Controllability.' Consistency ensures temporal and cross-view coherence, while Controllability ensures the alignment of generated content with corresponding annotations. Our approach integrates a novel 4D attention and a two-stage generation pipeline to maintain coherence, supplemented by the ControlNet framework for meticulous control by the Bird's-Eye-View (BEV) layouts. Extensive qualitative and quantitative evaluations of Panacea on the nuScenes dataset prove its effectiveness in generating high-quality multi-view driving-scene videos. This work notably propels the field of autonomous driving by effectively augmenting the training dataset used for advanced BEV perception techniques.
[](https://arxiv.org/abs/2311.16813) [](https://panacea-ad.github.io/) [](https://github.com/wenyuqing/panacea)
- **GeoDiffusion: Text-Prompted Geometric Control for Object Detection Data Generation**.
Abstract
Diffusion models have attracted significant attention due to the remarkable ability to create content and generate data for tasks like image classification. However, the usage of diffusion models to generate the high-quality object detection data remains an underexplored area, where not only image-level perceptual quality but also geometric conditions such as bounding boxes and camera views are essential. Previous studies have utilized either copy-paste synthesis or layout-to-image (L2I) generation with specifically designed modules to encode semantic layouts. In this paper, we propose GeoDiffusion, a simple framework that can flexibly translate various geometric conditions into text prompts and empower pre-trained text-to-image (T2I) diffusion models for high-quality detection data generation. Unlike previous L2I methods, our GeoDiffusion is able to encode not only the bounding boxes but also extra geometric conditions such as camera views in self-driving scenes. Extensive experiments demonstrate GeoDiffusion outperforms previous L2I methods while maintaining 4x training time faster. To the best of our knowledge, this is the first work to adopt diffusion models for layout-to-image generation with geometric conditions and demonstrate that L2I-generated images can be beneficial for improving the performance of object detectors.
[](https://arxiv.org/abs/2306.04607) [](https://kaichen1998.github.io/projects/geodiffusion/) [](https://github.com/KaiChen1998/GeoDiffusion/tree/main)
- **DrivingDiffusion: Layout-Guided multi-view driving scene video generation with latent diffusion model**.
Abstract
With the increasing popularity of autonomous driving based on the powerful and unified bird's-eye-view (BEV) representation, a demand for high-quality and large-scale multi-view video data with accurate annotation is urgently required. However, such large-scale multi-view data is hard to obtain due to expensive collection and annotation costs. To alleviate the problem, we propose a spatial-temporal consistent diffusion framework DrivingDiffusion, to generate realistic multi-view videos controlled by 3D layout. There are three challenges when synthesizing multi-view videos given a 3D layout: How to keep 1) cross-view consistency and 2) cross-frame consistency? 3) How to guarantee the quality of the generated instances? Our DrivingDiffusion solves the problem by cascading the multi-view single-frame image generation step, the single-view video generation step shared by multiple cameras, and post-processing that can handle long video generation. In the multi-view model, the consistency of multi-view images is ensured by information exchange between adjacent cameras. In the temporal model, we mainly query the information that needs attention in subsequent frame generation from the multi-view images of the first frame. We also introduce the local prompt to effectively improve the quality of generated instances. In post-processing, we further enhance the cross-view consistency of subsequent frames and extend the video length by employing temporal sliding window algorithm. Without any extra cost, our model can generate large-scale realistic multi-camera driving videos in complex urban scenes, fueling the downstream driving tasks. The code will be made publicly available.
[](https://arxiv.org/abs/2310.07771) [](https://drivingdiffusion.github.io/) [](https://github.com/shalfun/DrivingDiffusion)
- **MagicDrive: Street View Generation with Diverse 3D Geometry Control**.
Abstract
Recent advancements in diffusion models have significantly enhanced the data synthesis with 2D control. Yet, precise 3D control in street view generation, crucial for 3D perception tasks, remains elusive. Specifically, utilizing Bird's-Eye View (BEV) as the primary condition often leads to challenges in geometry control (e.g., height), affecting the representation of object shapes, occlusion patterns, and road surface elevations, all of which are essential to perception data synthesis, especially for 3D object detection tasks. In this paper, we introduce MagicDrive, a novel street view generation framework offering diverse 3D geometry controls, including camera poses, road maps, and 3D bounding boxes, together with textual descriptions, achieved through tailored encoding strategies. Besides, our design incorporates a cross-view attention module, ensuring consistency across multiple camera views. With MagicDrive, we achieve high-fidelity street-view synthesis that captures nuanced 3D geometry and various scene descriptions, enhancing tasks like BEV segmentation and 3D object detection.
[](https://arxiv.org/abs/2310.02601) [](https://flymin.github.io/magicdrive) [](https://github.com/cure-lab/MagicDrive)
- **DatasetDM: Synthesizing Data with Perception Annotations Using Diffusion Models**.
Abstract
Current deep networks are very data-hungry and benefit from training on largescale datasets, which are often time-consuming to collect and annotate. By contrast, synthetic data can be generated infinitely using generative models such as DALL-E and diffusion models, with minimal effort and cost. In this paper, we present DatasetDM, a generic dataset generation model that can produce diverse synthetic images and the corresponding high-quality perception annotations (e.g., segmentation masks, and depth). Our method builds upon the pre-trained diffusion model and extends text-guided image synthesis to perception data generation. We show that the rich latent code of the diffusion model can be effectively decoded as accurate perception annotations using a decoder module. Training the decoder only needs less than 1% (around 100 images) manually labeled images, enabling the generation of an infinitely large annotated dataset. Then these synthetic data can be used for training various perception models for downstream tasks. To showcase the power of the proposed approach, we generate datasets with rich dense pixel-wise labels for a wide range of downstream tasks, including semantic segmentation, instance segmentation, and depth estimation. Notably, it achieves 1) state-of-the-art results on semantic segmentation and instance segmentation; 2) significantly more robust on domain generalization than using the real data alone; and state-of-the-art results in zero-shot segmentation setting; and 3) flexibility for efficient application and novel task composition (e.g., image editing).
[](https://arxiv.org/abs/2308.06160) [](https://weijiawu.github.io/DatasetDM_page/) [](https://github.com/showlab/DatasetDM)
👆 [Back to Top](#Table-of-Content)
### NeRF
- **Neural Rendering based Urban Scene Reconstruction for Autonomous Driving**.
Abstract
Dense 3D reconstruction has many applications in automated driving including automated annotation validation, multimodal data augmentation, providing ground truth annotations for systems lacking LiDAR, as well as enhancing auto-labeling accuracy. LiDAR provides highly accurate but sparse depth, whereas camera images enable estimation of dense depth but noisy particularly at long ranges. In this paper, we harness the strengths of both sensors and propose a multimodal 3D scene reconstruction using a framework combining neural implicit surfaces and radiance fields. In particular, our method estimates dense and accurate 3D structures and creates an implicit map representation based on signed distance fields, which can be further rendered into RGB images, and depth maps. A mesh can be extracted from the learned signed distance field and culled based on occlusion. Dynamic objects are efficiently filtered on the fly during sampling using 3D object detection models. We demonstrate qualitative and quantitative results on challenging automotive scenes.
[](https://arxiv.org/abs/2402.06826)
- **OASim: an Open and Adaptive Simulator based on
Neural Rendering for Autonomous Driving**.
Abstract
With deep learning and computer vision technology development, autonomous driving provides new solutions to improve traffic safety and efficiency. The importance of building high-quality datasets is self-evident, especially with the rise of end-to-end autonomous driving algorithms in recent years. Data plays a core role in the algorithm closed-loop system. However, collecting real-world data is expensive, time-consuming, and unsafe. With the development of implicit rendering technology and in-depth research on using generative models to produce data at scale, we propose OASim, an open and adaptive simulator and autonomous driving data generator based on implicit neural rendering. It has the following characteristics: (1) High-quality scene reconstruction through neural implicit surface reconstruction technology. (2) Trajectory editing of the ego vehicle and participating vehicles. (3) Rich vehicle model library that can be freely selected and inserted into the scene. (4) Rich sensors model library where you can select specified sensors to generate data. (5) A highly customizable data generation system can generate data according to user needs. We demonstrate the high quality and fidelity of the generated data through perception performance evaluation on the Carla simulator and real-world data acquisition.
[](https://arxiv.org/abs/2402.03830) [](https://github.com/PJLab-ADG/OASim)
- **EmerNeRF: Emergent Spatial-Temporal Scene Decomposition via Self-Supervision**.
Abstract
We present EmerNeRF, a simple yet powerful approach for learning spatial-temporal representations of dynamic driving scenes. Grounded in neural fields, EmerNeRF simultaneously captures scene geometry, appearance, motion, and semantics via self-bootstrapping. EmerNeRF hinges upon two core components: First, it stratifies scenes into static and dynamic fields. This decomposition emerges purely from self-supervision, enabling our model to learn from general, in-the-wild data sources. Second, EmerNeRF parameterizes an induced flow field from the dynamic field and uses this flow field to further aggregate multi-frame features, amplifying the rendering precision of dynamic objects. Coupling these three fields (static, dynamic, and flow) enables EmerNeRF to represent highly-dynamic scenes self-sufficiently, without relying on ground truth object annotations or pre-trained models for dynamic object segmentation or optical flow estimation. Our method achieves state-of-the-art performance in sensor simulation, significantly outperforming previous methods when reconstructing static (+2.93 PSNR) and dynamic (+3.70 PSNR) scenes. In addition, to bolster EmerNeRF's semantic generalization, we lift 2D visual foundation model features into 4D space-time and address a general positional bias in modern Transformers, significantly boosting 3D perception performance (e.g., 37.50% relative improvement in occupancy prediction accuracy on average). Finally, we construct a diverse and challenging 120-sequence dataset to benchmark neural fields under extreme and highly-dynamic settings.
[](https://arxiv.org/abs/2311.02077) [](https://emernerf.github.io/) [](https://github.com/NVlabs/EmerNeRF)
- **UniSim: Synthesizing Data with Perception Annotations Using Diffusion Models**.
Abstract
Rigorously testing autonomy systems is essential for making safe self-driving vehicles (SDV) a reality. It requires one to generate safety critical scenarios beyond what can be collected safely in the world, as many scenarios happen rarely on public roads. To accurately evaluate performance, we need to test the SDV on these scenarios in closed-loop, where the SDV and other actors interact with each other at each timestep. Previously recorded driving logs provide a rich resource to build these new scenarios from, but for closed loop evaluation, we need to modify the sensor data based on the new scene configuration and the SDV's decisions, as actors might be added or removed and the trajectories of existing actors and the SDV will differ from the original log. In this paper, we present UniSim, a neural sensor simulator that takes a single recorded log captured by a sensor-equipped vehicle and converts it into a realistic closed-loop multi-sensor simulation. UniSim builds neural feature grids to reconstruct both the static background and dynamic actors in the scene, and composites them together to simulate LiDAR and camera data at new viewpoints, with actors added or removed and at new placements. To better handle extrapolated views, we incorporate learnable priors for dynamic objects, and leverage a convolutional network to complete unseen regions. Our experiments show UniSim can simulate realistic sensor data with small domain gap on downstream tasks. With UniSim, we demonstrate closed-loop evaluation of an autonomy system on safety-critical scenarios as if it were in the real world.
[](https://arxiv.org/abs/2308.01898) [](https://waabi.ai/unisim/)
- **MARS: An Instance-aware, Modular and Realistic Simulator for Autonomous Driving**.
Abstract
Nowadays, autonomous cars can drive smoothly in ordinary cases, and it is widely recognized that realistic sensor simulation will play a critical role in solving remaining corner cases by simulating them. To this end, we propose an autonomous driving simulator based upon neural radiance fields (NeRFs). Compared with existing works, ours has three notable features: (1) Instance-aware. Our simulator models the foreground instances and background environments separately with independent networks so that the static (e.g., size and appearance) and dynamic (e.g., trajectory) properties of instances can be controlled separately. (2) Modular. Our simulator allows flexible switching between different modern NeRF-related backbones, sampling strategies, input modalities, etc. We expect this modular design to boost academic progress and industrial deployment of NeRF-based autonomous driving simulation. (3) Realistic. Our simulator set new state-of-the-art photo-realism results given the best module selection. Our simulator will be open-sourced while most of our counterparts are not.
[](https://arxiv.org/abs/2307.15058) [](https://open-air-sun.github.io/mars/) [](https://github.com/OPEN-AIR-SUN/mars)
- **NeRF-LiDAR: Generating Realistic LiDAR Point Clouds with Neural Radiance Fields**.
Abstract
Labeling LiDAR point clouds for training autonomous driving is extremely expensive and difficult. LiDAR simulation aims at generating realistic LiDAR data with labels for training and verifying self-driving algorithms more efficiently. Recently, Neural Radiance Fields (NeRF) have been proposed for novel view synthesis using implicit reconstruction of 3D scenes. Inspired by this, we present NeRF-LIDAR, a novel LiDAR simulation method that leverages real-world information to generate realistic LIDAR point clouds. Different from existing LiDAR simulators, we use real images and point cloud data collected by self-driving cars to learn the 3D scene representation, point cloud generation and label rendering. We verify the effectiveness of our NeRF-LiDAR by training different 3D segmentation models on the generated LiDAR point clouds. It reveals that the trained models are able to achieve similar accuracy when compared with the same model trained on the real LiDAR data. Besides, the generated data is capable of boosting the accuracy through pre-training which helps reduce the requirements of the real labeled data.
[](https://arxiv.org/abs/2304.14811)
- **StreetSurf: Extending Multi-view Implicit Surface Reconstruction to Street Views**.
Abstract
We present a novel multi-view implicit surface reconstruction technique, termed StreetSurf, that is readily applicable to street view images in widely-used autonomous driving datasets, such as Waymo-perception sequences, without necessarily requiring LiDAR data. As neural rendering research expands rapidly, its integration into street views has started to draw interests. Existing approaches on street views either mainly focus on novel view synthesis with little exploration of the scene geometry, or rely heavily on dense LiDAR data when investigating reconstruction. Neither of them investigates multi-view implicit surface reconstruction, especially under settings without LiDAR data. Our method extends prior object-centric neural surface reconstruction techniques to address the unique challenges posed by the unbounded street views that are captured with non-object-centric, long and narrow camera trajectories. We delimit the unbounded space into three parts, close-range, distant-view and sky, with aligned cuboid boundaries, and adapt cuboid/hyper-cuboid hash-grids along with road-surface initialization scheme for finer and disentangled representation. To further address the geometric errors arising from textureless regions and insufficient viewing angles, we adopt geometric priors that are estimated using general purpose monocular models. Coupled with our implementation of efficient and fine-grained multi-stage ray marching strategy, we achieve state of the art reconstruction quality in both geometry and appearance within only one to two hours of training time with a single RTX3090 GPU for each street view sequence. Furthermore, we demonstrate that the reconstructed implicit surfaces have rich potential for various downstream tasks, including ray tracing and LiDAR simulation.
[](https://arxiv.org/abs/2306.04988) [](https://ventusff.github.io/streetsurf_web/) [](https://github.com/pjlab-ADG/neuralsim)
- **MapNeRF: Incorporating Map Priors into Neural Radiance Fields for Driving View Simulation**.
Abstract
Simulating camera sensors is a crucial task in autonomous driving. Although neural radiance fields are exceptional at synthesizing photorealistic views in driving simulations, they still fail to generate extrapolated views. This paper proposes to incorporate map priors into neural radiance fields to synthesize out-of-trajectory driving views with semantic road consistency. The key insight is that map information can be utilized as a prior to guiding the training of the radiance fields with uncertainty. Specifically, we utilize the coarse ground surface as uncertain information to supervise the density field and warp depth with uncertainty from unknown camera poses to ensure multi-view consistency. Experimental results demonstrate that our approach can produce semantic consistency in deviated views for vehicle camera simulation.
[](https://arxiv.org/abs/2307.14981) [](https://www.youtube.com/watch?v=jEQWr-Rfh3A&feature=youtu.be)
- **RoMe: Towards Large Scale Road Surface Reconstruction via Mesh Representation**.
Abstract
In autonomous driving applications, accurate and efficient road surface reconstruction is paramount. This paper introduces RoMe, a novel framework designed for the robust reconstruction of large-scale road surfaces. Leveraging a unique mesh representation, RoMe ensures that the reconstructed road surfaces are accurate and seamlessly aligned with semantics. To address challenges in computational efficiency, we propose a waypoint sampling strategy, enabling RoMe to reconstruct vast environments by focusing on sub-areas and subsequently merging them. Furthermore, we incorporate an extrinsic optimization module to enhance the robustness against inaccuracies in extrinsic calibration. Our extensive evaluations of both public datasets and wild data underscore RoMe's superiority in terms of speed, accuracy, and robustness. For instance, it costs only 2 GPU hours to recover a road surface of 600*600 square meters from thousands of images. Notably, RoMe's capability extends beyond mere reconstruction, offering significant value for auto-labeling tasks in autonomous driving applications.
[](https://arxiv.org/abs/2306.11368) [](https://github.com/DRosemei/RoMe)
- **S-NeRF: Neural Radiance Fields for Street Views**.
Abstract
Neural Radiance Fields (NeRFs) aim to synthesize novel views of objects and scenes, given the object-centric camera views with large overlaps. However, we conjugate that this paradigm does not fit the nature of the street views that are collected by many self-driving cars from the large-scale unbounded scenes. Also, the onboard cameras perceive scenes without much overlapping. Thus, existing NeRFs often produce blurs, 'floaters' and other artifacts on street-view synthesis. In this paper, we propose a new street-view NeRF (S-NeRF) that considers novel view synthesis of both the large-scale background scenes and the foreground moving vehicles jointly. Specifically, we improve the scene parameterization function and the camera poses for learning better neural representations from street views. We also use the the noisy and sparse LiDAR points to boost the training and learn a robust geometry and reprojection based confidence to address the depth outliers. Moreover, we extend our S-NeRF for reconstructing moving vehicles that is impracticable for conventional NeRFs. Thorough experiments on the large-scale driving datasets (e.g., nuScenes and Waymo) demonstrate that our method beats the state-of-the-art rivals by reducing 7% to 40% of the mean-squared error in the street-view synthesis and a 45% PSNR gain for the moving vehicles rendering.
[](https://arxiv.org/abs/2303.00749) [](https://ziyang-xie.github.io/s-nerf) [](https://github.com/fudan-zvg/S-NeRF)
- **LiDAR-NeRF: Novel LiDAR View Synthesis via Neural Radiance Fields**.
Abstract
We introduce a new task, novel view synthesis for LiDAR sensors. While traditional model-based LiDAR simulators with style-transfer neural networks can be applied to render novel views, they fall short of producing accurate and realistic LiDAR patterns because the renderers rely on explicit 3D reconstruction and exploit game engines, that ignore important attributes of LiDAR points. We address this challenge by formulating, to the best of our knowledge, the first differentiable end-to-end LiDAR rendering framework, LiDAR-NeRF, leveraging a neural radiance field (NeRF) to facilitate the joint learning of geometry and the attributes of 3D points. However, simply employing NeRF cannot achieve satisfactory results, as it only focuses on learning individual pixels while ignoring local information, especially at low texture areas, resulting in poor geometry. To this end, we have taken steps to address this issue by introducing a structural regularization method to preserve local structural details. To evaluate the effectiveness of our approach, we establish an object-centric multi-view LiDAR dataset, dubbed NeRF-MVL. It contains observations of objects from 9 categories seen from 360-degree viewpoints captured with multiple LiDAR sensors. Our extensive experiments on the scene-level KITTI-360 dataset, and on our object-level NeRF-MVL show that our LiDAR-NeRF surpasses the model-based algorithms significantly.
[](https://arxiv.org/abs/2304.10406) [](https://tangtaogo.github.io/lidar-nerf-website/) [](https://github.com/tangtaogo/lidar-nerf)
- **READ: Large-Scale Neural Scene Rendering for Autonomous Driving**.
Abstract
Synthesizing free-view photo-realistic images is an important task in multimedia. With the development of advanced driver assistance systems~(ADAS) and their applications in autonomous vehicles, experimenting with different scenarios becomes a challenge. Although the photo-realistic street scenes can be synthesized by image-to-image translation methods, which cannot produce coherent scenes due to the lack of 3D information. In this paper, a large-scale neural rendering method is proposed to synthesize the autonomous driving scene~(READ), which makes it possible to synthesize large-scale driving scenarios on a PC through a variety of sampling schemes. In order to represent driving scenarios, we propose an {\omega} rendering network to learn neural descriptors from sparse point clouds. Our model can not only synthesize realistic driving scenes but also stitch and edit driving scenes. Experiments show that our model performs well in large-scale driving scenarios.
[](https://arxiv.org/abs/2205.05509) [](https://github.com/JOP-Lee/READ)
👆 [Back to Top](#Table-of-Content)
### 3D Gaussian Splatting
- **TCLC-GS: Tightly Coupled LiDAR-Camera Gaussian Splatting for Surrounding Autonomous Driving Scenes**.
Abstract
Most 3D Gaussian Splatting (3D-GS) based methods for urban scenes initialize 3D Gaussians directly with 3D LiDAR points, which not only underutilizes LiDAR data capabilities but also overlooks the potential advantages of fusing LiDAR with camera data. In this paper, we design a novel tightly coupled LiDAR-Camera Gaussian Splatting (TCLC-GS) to fully leverage the combined strengths of both LiDAR and camera sensors, enabling rapid, high-quality 3D reconstruction and novel view RGB/depth synthesis. TCLC-GS designs a hybrid explicit (colorized 3D mesh) and implicit (hierarchical octree feature) 3D representation derived from LiDAR-camera data, to enrich the properties of 3D Gaussians for splatting. 3D Gaussian's properties are not only initialized in alignment with the 3D mesh which provides more completed 3D shape and color information, but are also endowed with broader contextual information through retrieved octree implicit features. During the Gaussian Splatting optimization process, the 3D mesh offers dense depth information as supervision, which enhances the training process by learning of a robust geometry. Comprehensive evaluations conducted on the Waymo Open Dataset and nuScenes Dataset validate our method's state-of-the-art (SOTA) performance. Utilizing a single NVIDIA RTX 3090 Ti, our method demonstrates fast training and achieves real-time RGB and depth rendering at 90 FPS in resolution of 1920x1280 (Waymo), and 120 FPS in resolution of 1600x900 (nuScenes) in urban scenarios.
[](https://arxiv.org/abs/2404.02410)
- **DC-Gaussian: Improving 3D Gaussian Splatting for Reflective Dash Cam Videos**.
Abstract
We present DC-Gaussian, a new method for generating novel views from in-vehicle dash cam videos.
While neural rendering techniques have made significant strides in driving scenarios, existing methods are primarily designed for videos collected by autonomous vehicles. However, these videos are limited in both quantity and diversity compared to dash cam videos, which are more widely used across various types of vehicles and capture a broader range of scenarios. Dash cam videos often suffer from severe obstructions such as reflections and occlusions on the windshields, which significantly impede the application of neural rendering techniques. To address this challenge, we develop DC-Gaussian based on the recent real-time neural rendering technique 3D Gaussian Splatting (3DGS). Our approach includes an adaptive image decomposition module to model reflections and occlusions in a unified manner. Additionally, we introduce illumination- aware obstruction modeling to manage reflections and occlusions under varying lighting conditions. Lastly, we employ a geometry-guided Gaussian enhancement strategy to improve rendering details by incorporating additional geometry priors.
Experiments on self-captured and public dash cam videos show that our method not only achieves state-of-the-art performance in novel view synthesis, but also accurately reconstructing captured scenes getting rid of obstructions.
[](https://arxiv.org/abs/2405.17705) [](https://linhanwang.github.io/dcgaussian/) [](https://github.com/linhanwang/DC-Gaussian)
- **Street Gaussians for Modeling Dynamic Urban Scenes**.
Abstract
This paper aims to tackle the problem of modeling dynamic urban street scenes from monocular videos. Recent methods extend NeRF by incorporating tracked vehicle poses to animate vehicles, enabling photo-realistic view synthesis of dynamic urban street scenes. However, significant limitations are their slow training and rendering speed, coupled with the critical need for high precision in tracked vehicle poses. We introduce Street Gaussians, a new explicit scene representation that tackles all these limitations. Specifically, the dynamic urban street is represented as a set of point clouds equipped with semantic logits and 3D Gaussians, each associated with either a foreground vehicle or the background. To model the dynamics of foreground object vehicles, each object point cloud is optimized with optimizable tracked poses, along with a dynamic spherical harmonics model for the dynamic appearance. The explicit representation allows easy composition of object vehicles and background, which in turn allows for scene editing operations and rendering at 133 FPS (1066×1600 resolution) within half an hour of training. The proposed method is evaluated on multiple challenging benchmarks, including KITTI and Waymo Open datasets. Experiments show that the proposed method consistently outperforms state-of-the-art methods across all datasets. Furthermore, the proposed representation delivers performance on par with that achieved using precise ground-truth poses, despite relying only on poses from an off-the-shelf tracker.
[](https://arxiv.org/abs/2401.01339) [](https://zju3dv.github.io/street_gaussians/) [](https://github.com/zju3dv/street_gaussians)
- **DrivingGaussian: Composite Gaussian Splatting for Surrounding Dynamic Autonomous Driving Scenes**.
Abstract
We present DrivingGaussian, an efficient and effective framework for surrounding dynamic autonomous driving scenes. For complex scenes with moving objects, we first sequentially and progressively model the static background of the entire scene with incremental static 3D Gaussians. We then leverage a composite dynamic Gaussian graph to handle multiple moving objects, individually reconstructing each object and restoring their accurate positions and occlusion relationships within the scene. We further use a LiDAR prior for Gaussian Splatting to reconstruct scenes with greater details and maintain panoramic consistency. DrivingGaussian outperforms existing methods in driving scene reconstruction and enables photorealistic surround-view synthesis with high-fidelity and multi-camera consistency.
[](https://arxiv.org/abs/2312.07920) [](https://pkuvdig.github.io/DrivingGaussian/)
- **Periodic Vibration Gaussian: Dynamic Urban Scene Reconstruction and Real-time Rendering**.
Abstract
Modeling dynamic, large-scale urban scenes is challenging due to their highly intricate geometric structures and unconstrained dynamics in both space and time. Prior methods often employ high-level architectural priors, separating static and dynamic elements, resulting in suboptimal capture of their synergistic interactions. To address this challenge, we present a unified representation model, called Periodic Vibration Gaussian (PVG). PVG builds upon the efficient 3D Gaussian splatting technique, originally designed for static scene representation, by introducing periodic vibration-based temporal dynamics. This innovation enables PVG to elegantly and uniformly represent the characteristics of various objects and elements in dynamic urban scenes. To enhance temporally coherent representation learning with sparse training data, we introduce a novel flow-based temporal smoothing mechanism and a position-aware adaptive control strategy. Extensive experiments on Waymo Open Dataset and KITTI benchmarks demonstrate that PVG surpasses state-of-the-art alternatives in both reconstruction and novel view synthesis for both dynamic and static scenes. Notably, PVG achieves this without relying on manually labeled object bounding boxes or expensive optical flow estimation. Moreover, PVG exhibits 50/6000-fold acceleration in training/rendering over the best alternative.
[](https://arxiv.org/abs/2311.18561) [](https://fudan-zvg.github.io/PVG/)
👆 [Back to Top](#Table-of-Content)
## Self-supervised Training
### Contrastive
A survey paper of contrastive-based self-supervised learning: [A survey on contrastive self-supervised learning](https://arxiv.org/abs/2011.00362).
- **BEVContrast: Self-Supervision in BEV Space for Automotive Lidar Point Clouds**.
Abstract
We present a surprisingly simple and efficient method for self-supervision of 3D backbone on automotive Lidar point clouds. We design a contrastive loss between features of Lidar scans captured in the same scene. Several such approaches have been proposed in the literature from PointConstrast [40 ], which uses a contrast at the level of points, to the state-of-the-art TARL [30 ], which uses a contrast at the level of segments, roughly corresponding to objects. While the former enjoys a great simplicity of implementation, it is surpassed by the latter, which however requires a costly pre-processing. In BEVContrast, we define our contrast at the level of 2D cells in the Bird's Eye View plane. Resulting cell-level representations offer a good trade-off between the point-level representations exploited in PointContrast and segment-level representations exploited in TARL: we retain the simplicity of PointContrast (cell representations are cheap to compute) while surpassing the performance of TARL in downstream semantic segmentation.
[](https://arxiv.org/abs/2310.17281) [](https://github.com/valeoai/BEVContrast)
- **AD-PT: Autonomous Driving Pre-Training with Large-scale Point Cloud Dataset**.
Abstract
It is a long-term vision for Autonomous Driving (AD) community that the perception models can learn from a large-scale point cloud dataset, to obtain unified representations that can achieve promising results on different tasks or benchmarks. Previous works mainly focus on the self-supervised pre-training pipeline, meaning that they perform the pre-training and fine-tuning on the same benchmark, which is difficult to attain the performance scalability and cross-dataset application for the pre-training checkpoint. In this paper, for the first time, we are committed to building a large-scale pre-training point-cloud dataset with diverse data distribution, and meanwhile learning generalizable representations from such a diverse pre-training dataset. We formulate the point-cloud pre-training task as a semi-supervised problem, which leverages the few-shot labeled and massive unlabeled point-cloud data to generate the unified backbone representations that can be directly applied to many baseline models and benchmarks, decoupling the AD-related pre-training process and downstream fine-tuning task. During the period of backbone pre-training, by enhancing the scene- and instance-level distribution diversity and exploiting the backbone's ability to learn from unknown instances, we achieve significant performance gains on a series of downstream perception benchmarks including Waymo, nuScenes, and KITTI, under different baseline models like PV-RCNN++, SECOND, CenterPoint.
[](https://arxiv.org/abs/2306.00612) [](https://jiakangyuan.github.io/AD-PT.github.io/) [](https://github.com/PJLab-ADG/3DTrans)
- **SegContrast: 3D Point Cloud Feature Representation Learning Through Self-Supervised Segment Discrimination**.
Abstract
Semantic scene interpretation is essential for autonomous systems to operate in complex scenarios. While deep learning-based methods excel at this task, they rely on vast amounts of labeled data that is tedious to generate and might not cover all relevant classes sufficiently. Self-supervised representation learning has the prospect of reducing the amount of required labeled data by learning descriptive representations from unlabeled data. In this letter, we address the problem of representation learning for 3D point cloud data in the context of autonomous driving. We propose a new contrastive learning approach that aims at learning the structural context of the scene. Our approach extracts class-agnostic segments over the point cloud and applies the contrastive loss over these segments to discriminate between similar and dissimilar structures. We apply our method on data recorded with a 3D LiDAR. We show that our method achieves competitive performance and can learn a more descriptive feature representation than other state-of-the-art self-supervised contrastive point cloud methods.
[](https://www.ipb.uni-bonn.de/pdfs/nunes2022ral-icra.pdf) [](https://github.com/PRBonn/segcontrast)
- **Temporal Consistent 3D LiDAR Representation Learning for Semantic Perception in Autonomous Driving**.
Abstract
Semantic perception is a core building block in autonomous driving, since it provides information about the drivable space and location of other traffic participants. For learning-based perception, often a large amount of diverse training data is necessary to achieve high performance. Data labeling is usually a bottleneck for developing such methods, especially for dense prediction tasks, e.g., semantic segmentation or panoptic segmentation. For 3D LiDAR data, the annotation process demands even more effort than for images. Especially in autonomous driving, point clouds are sparse, and objects appearance depends on its distance from the sensor, making it harder to acquire large amounts of labeled training data. This paper aims at taking an alternative path proposing a self-supervised representation learning method for 3D LiDAR data. Our approach exploits the vehicle motion to match objects across time viewed in different scans. We then train a model to maximize the point-wise feature similarities from points of the associated object in different scans, which enables to learn a consistent representation across time. The experimental results show that our approach performs better than previous state-of-the-art self-supervised representation learning methods when fine-tuning to different downstream tasks. We furthermore show that with only 10% of labeled data, a network pre-trained with our approach can achieve better performance than the same network trained from scratch with all labels for semantic segmentation on SemanticKITTI.
[](https://openaccess.thecvf.com/content/CVPR2023/html/Nunes_Temporal_Consistent_3D_LiDAR_Representation_Learning_for_Semantic_Perception_in_CVPR_2023_paper.html) [](https://github.com/PRBonn/TARL)
- **SimIPU: Simple 2D Image and 3D Point Cloud Unsupervised Pre-Training for Spatial-Aware Visual Representations**.
Abstract
Pre-training has become a standard paradigm in many computer vision tasks. However, most of the methods are generally designed on the RGB image domain. Due to the discrepancy between the two-dimensional image plane and the three-dimensional space, such pre-trained models fail to perceive spatial information and serve as sub-optimal solutions for 3D-related tasks. To bridge this gap, we aim to learn a spatial-aware visual representation that can describe the three-dimensional space and is more suitable and effective for these tasks. To leverage point clouds, which are much more superior in providing spatial information compared to images, we propose a simple yet effective 2D Image and 3D Point cloud Unsupervised pre-training strategy, called SimIPU. Specifically, we develop a multi-modal contrastive learning framework that consists of an intra-modal spatial perception module to learn a spatial-aware representation from point clouds and an inter-modal feature interaction module to transfer the capability of perceiving spatial information from the point cloud encoder to the image encoder, respectively. Positive pairs for contrastive losses are established by the matching algorithm and the projection matrix. The whole framework is trained in an unsupervised end-to-end fashion. To the best of our knowledge, this is the first study to explore contrastive learning pre-training strategies for outdoor multi-modal datasets, containing paired camera images and LIDAR point clouds.
[](https://arxiv.org/abs/2112.04680) [](https://github.com/zhyever/SimIPU)
- **Self-Supervised Pretraining of 3D Features on any Point-Cloud**.
Abstract
Pretraining on large labeled datasets is a prerequisite to achieve good performance in many computer vision tasks like 2D object recognition, video classification etc. However, pretraining is not widely used for 3D recognition tasks where state-of-the-art methods train models from scratch. A primary reason is the lack of large annotated datasets because 3D data is both difficult to acquire and time consuming to label. We present a simple self-supervised pertaining method that can work with any 3D data - single or multiview, indoor or outdoor, acquired by varied sensors, without 3D registration. We pretrain standard point cloud and voxel based model architectures, and show that joint pretraining further improves performance. We evaluate our models on 9 benchmarks for object detection, semantic segmentation, and object classification, where they achieve state-of-the-art results and can outperform supervised pretraining. We set a new state-of-the-art for object detection on ScanNet (69.0% mAP) and SUNRGBD (63.5% mAP).
[](https://arxiv.org/abs/2101.02691) [](https://github.com/facebookresearch/DepthContrast)
- **PointContrast: Unsupervised Pre-training for 3D Point Cloud Understanding**.
Abstract
Arguably one of the top success stories of deep learning is transfer learning. The finding that pre-training a network on a rich source set (e.g., ImageNet) can help boost performance once fine-tuned on a usually much smaller target set, has been instrumental to many applications in language and vision. Yet, very little is known about its usefulness in 3D point cloud understanding. We see this as an opportunity considering the effort required for annotating data in 3D. In this work, we aim at facilitating research on 3D representation learning. Different from previous works, we focus on high-level scene understanding tasks. To this end, we select a suite of diverse datasets and tasks to measure the effect of unsupervised pre-training on a large source set of 3D scenes. Our findings are extremely encouraging: using a unified triplet of architecture, source dataset, and contrastive loss for pre-training, we achieve improvement over recent best results in segmentation and detection across 6 different benchmarks for indoor and outdoor, real and synthetic datasets – demonstrating that the learned representation can generalize across domains. Furthermore, the improvement was similar to supervised pre-training, suggesting that future efforts should favor scaling data collection over more detailed annotation. We hope these findings will encourage more research on unsupervised pretext task design for 3D deep learning.
[](https://arxiv.org/abs/2007.10985) [](https://github.com/facebookresearch/PointContrast)
👆 [Back to Top](#Table-of-Content)
### Reconstruction
- **Voxel-MAE - Masked Autoencoders for Self-Supervised Learning on Automotive Point Clouds**.
Abstract
Masked autoencoding has become a successful pretraining paradigm for Transformer models for text, images, and, recently, point clouds. Raw automotive datasets are suitable candidates for self-supervised pre-training as they generally are cheap to collect compared to annotations for tasks like 3D object detection (OD). However, the development of masked autoencoders for point clouds has focused solely on synthetic and indoor data. Consequently, existing methods have tailored their representations and models toward small and dense point clouds with homogeneous point densities. In this work, we study masked autoencoding for point clouds in an automotive setting, which are sparse and for which the point density can vary drastically among objects in the same scene. To this end, we propose Voxel-MAE, a simple masked autoencoding pre-training scheme designed for voxel representations. We pre-train the backbone of a Transformer-based 3D object detector to reconstruct masked voxels and to distinguish between empty and non-empty voxels. Our method improves the 3D OD performance by 1.75 mAP points and 1.05 NDS on the challenging nuScenes dataset. Further, we show that by pre-training with Voxel-MAE, we require only 40% of the annotated data to outperform a randomly initialized equivalent.
[](https://arxiv.org/abs/2207.00531) [](https://georghess.se/projects/voxel-mae/) [](https://github.com/georghess/voxel-mae)
- **GD-MAE: Generative Decoder for MAE Pre-training on LiDAR Point Clouds**.
Abstract
Despite the tremendous progress of Masked Autoencoders (MAE) in developing vision tasks such as image and video, exploring MAE in large-scale 3D point clouds remains challenging due to the inherent irregularity. In contrast to previous 3D MAE frameworks, which either design a complex decoder to infer masked information from maintained regions or adopt sophisticated masking strategies, we instead propose a much simpler paradigm. The core idea is to apply a \textbf{G}enerative \textbf{D}ecoder for MAE (GD-MAE) to automatically merges the surrounding context to restore the masked geometric knowledge in a hierarchical fusion manner. In doing so, our approach is free from introducing the heuristic design of decoders and enjoys the flexibility of exploring various masking strategies. The corresponding part costs less than \textbf{12\%} latency compared with conventional methods, while achieving better performance. We demonstrate the efficacy of the proposed method on several large-scale benchmarks: Waymo, KITTI, and ONCE. Consistent improvement on downstream detection tasks illustrates strong robustness and generalization capability. Not only our method reveals state-of-the-art results, but remarkably, we achieve comparable accuracy even with \textbf{20\%} of the labeled data on the Waymo dataset.
[](https://arxiv.org/abs/2212.03010) [](https://github.com/Nightmare-n/GD-MAE)
- **UniM2AE: Multi-modal Masked Autoencoders with Unified 3D Representation for 3D Perception in Autonomous Driving**.
Abstract
Masked Autoencoders (MAE) play a pivotal role in learning potent representations, delivering outstanding results across various 3D perception tasks essential for autonomous driving. In real-world driving scenarios, it's commonplace to deploy multiple sensors for comprehensive environment perception. While integrating multi-modal features from these sensors can produce rich and powerful features, there is a noticeable gap in MAE methods addressing this integration. This research delves into multi-modal Masked Autoencoders tailored for a unified representation space in autonomous driving, aiming to pioneer a more efficient fusion of two distinct modalities. To intricately marry the semantics inherent in images with the geometric intricacies of LiDAR point clouds, the UniM$^2$AE is proposed. This model stands as a potent yet straightforward, multi-modal self-supervised pre-training framework, mainly consisting of two designs. First, it projects the features from both modalities into a cohesive 3D volume space, ingeniously expanded from the bird's eye view (BEV) to include the height dimension. The extension makes it possible to back-project the informative features, obtained by fusing features from both modalities, into their native modalities to reconstruct the multiple masked inputs. Second, the Multi-modal 3D Interactive Module (MMIM) is invoked to facilitate the efficient inter-modal interaction during the interaction process. Extensive experiments conducted on the nuScenes Dataset attest to the efficacy of UniM$^2$AE, indicating enhancements in 3D object detection and BEV map segmentation by 1.2\%(NDS) and 6.5\% (mIoU), respectively.
[](https://arxiv.org/abs/2308.10421) [](https://github.com/hollow-503/UniM2AE)
- **UniScene: Multi-Camera Unified Pre-training via 3D Scene Reconstruction**.
Abstract
Multi-camera 3D perception has emerged as a prominent research field in autonomous driving, offering a viable and cost-effective alternative to LiDAR-based solutions. The existing multi-camera algorithms primarily rely on monocular 2D pre-training. However, the monocular 2D pre-training overlooks the spatial and temporal correlations among the multi-camera system. To address this limitation, we propose the first multi-camera unified pre-training framework, called UniScene, which involves initially reconstructing the 3D scene as the foundational stage and subsequently fine-tuning the model on downstream tasks. Specifically, we employ Occupancy as the general representation for the 3D scene, enabling the model to grasp geometric priors of the surrounding world through pre-training. A significant benefit of UniScene is its capability to utilize a considerable volume of unlabeled image-LiDAR pairs for pre-training purposes. The proposed multi-camera unified pre-training framework demonstrates promising results in key tasks such as multi-camera 3D object detection and surrounding semantic scene completion. When compared to monocular pre-training methods on the nuScenes dataset, UniScene shows a significant improvement of about 2.0% in mAP and 2.0% in NDS for multi-camera 3D object detection, as well as a 3% increase in mIoU for surrounding semantic scene completion. By adopting our unified pre-training method, a 25% reduction in 3D training annotation costs can be achieved, offering significant practical value for the implementation of real-world autonomous driving.
[](https://arxiv.org/abs/2305.18829) [](https://github.com/chaytonmin/UniScene)
- **Occupancy-MAE: Self-supervised Pre-training Large-scale LiDAR Point Clouds with Masked Occupancy Autoencoders**.
Abstract
Current perception models in autonomous driving heavily rely on large-scale labelled 3D data, which is both costly and time-consuming to annotate. This work proposes a solution to reduce the dependence on labelled 3D training data by leveraging pre-training on large-scale unlabeled outdoor LiDAR point clouds using masked autoencoders (MAE). While existing masked point autoencoding methods mainly focus on small-scale indoor point clouds or pillar-based large-scale outdoor LiDAR data, our approach introduces a new self-supervised masked occupancy pre-training method called Occupancy-MAE, specifically designed for voxel-based large-scale outdoor LiDAR point clouds. Occupancy-MAE takes advantage of the gradually sparse voxel occupancy structure of outdoor LiDAR point clouds and incorporates a range-aware random masking strategy and a pretext task of occupancy prediction. By randomly masking voxels based on their distance to the LiDAR and predicting the masked occupancy structure of the entire 3D surrounding scene, Occupancy-MAE encourages the extraction of high-level semantic information to reconstruct the masked voxel using only a small number of visible voxels. Extensive experiments demonstrate the effectiveness of Occupancy-MAE across several downstream tasks. For 3D object detection, Occupancy-MAE reduces the labelled data required for car detection on the KITTI dataset by half and improves small object detection by approximately 2% in AP on the Waymo dataset. For 3D semantic segmentation, Occupancy-MAE outperforms training from scratch by around 2% in mIoU. For multi-object tracking, Occupancy-MAE enhances training from scratch by approximately 1% in terms of AMOTA and AMOTP.
[](https://arxiv.org/abs/2206.09900) [](https://github.com/chaytonmin/Occupancy-MAE)
- **Implicit Autoencoder for Point-Cloud Self-Supervised Representation Learning**.
Abstract
This paper advocates the use of implicit surface representation in autoencoder-based self-supervised 3D representation learning. The most popular and accessible 3D representation, i.e., point clouds, involves discrete samples of the underlying continuous 3D surface. This discretization process introduces sampling variations on the 3D shape, making it challenging to develop transferable knowledge of the true 3D geometry. In the standard autoencoding paradigm, the encoder is compelled to encode not only the 3D geometry but also information on the specific discrete sampling of the 3D shape into the latent code. This is because the point cloud reconstructed by the decoder is considered unacceptable unless there is a perfect mapping between the original and the reconstructed point clouds. This paper introduces the Implicit AutoEncoder (IAE), a simple yet effective method that addresses the sampling variation issue by replacing the commonly-used point-cloud decoder with an implicit decoder. The implicit decoder reconstructs a continuous representation of the 3D shape, independent of the imperfections in the discrete samples. Extensive experiments demonstrate that the proposed IAE achieves state-of-the-art performance across various self-supervised learning benchmarks.
[](https://arxiv.org/abs/2201.00785) [](https://github.com/SimingYan/IAE)
- **BEV-MAE: Bird's Eye View Masked Autoencoders for Outdoor Point Cloud Pre-training**.
Abstract
Current outdoor LiDAR-based 3D object detection methods mainly adopt the training-from-scratch paradigm. Unfortunately, this paradigm heavily relies on large-scale labeled data, whose collection can be expensive and time-consuming. Self-supervised pre-training is an effective and desirable way to alleviate this dependence on extensive annotated data. Recently, masked modeling has become a successful self-supervised learning approach for point clouds. However, current works mainly focus on synthetic or indoor datasets. When applied to large-scale and sparse outdoor point clouds, they fail to yield satisfactory results. In this work, we present BEV-MAE, a simple masked autoencoder pre-training framework for 3D object detection on outdoor point clouds. Specifically, we first propose a bird's eye view (BEV) guided masking strategy to guide the 3D encoder learning feature representation in a BEV perspective and avoid complex decoder design during pre-training. Besides, we introduce a learnable point token to maintain a consistent receptive field size of the 3D encoder with fine-tuning for masked point cloud inputs. Finally, based on the property of outdoor point clouds, i.e., the point clouds of distant objects are more sparse, we propose point density prediction to enable the 3D encoder to learn location information, which is essential for object detection. Experimental results show that BEV-MAE achieves new state-of-the-art self-supervised results on both Waymo and nuScenes with diverse 3D object detectors. Furthermore, with only 20% data and 7% training cost during pre-training, BEV-MAE achieves comparable performance with the state-of-the-art method ProposalContrast.
[](https://arxiv.org/abs/2212.05758)
- **MAELi: Masked Autoencoder for Large-Scale LiDAR Point Clouds**.
Abstract
The sensing process of large-scale LiDAR point clouds inevitably causes large blind spots, i.e. regions not visible to the sensor. We demonstrate how these inherent sampling properties can be effectively utilized for self-supervised representation learning by designing a highly effective pre-training framework that considerably reduces the need for tedious 3D annotations to train state-of-the-art object detectors. Our Masked AutoEncoder for LiDAR point clouds (MAELi) intuitively leverages the sparsity of LiDAR point clouds in both the encoder and decoder during reconstruction. This results in more expressive and useful initialization, which can be directly applied to downstream perception tasks, such as 3D object detection or semantic segmentation for autonomous driving. In a novel reconstruction approach, MAELi distinguishes between empty and occluded space and employs a new masking strategy that targets the LiDAR's inherent spherical projection. Thereby, without any ground truth whatsoever and trained on single frames only, MAELi obtains an understanding of the underlying 3D scene geometry and semantics. To demonstrate the potential of MAELi, we pre-train backbones in an end-to-end manner and show the effectiveness of our unsupervised pre-trained weights on the tasks of 3D object detection and semantic segmentation.
[](https://arxiv.org/abs/2212.07207)
- **GD-MAE: Generative Decoder for MAE Pre-training on LiDAR Point Clouds**.
Abstract
Despite the tremendous progress of Masked Autoencoders (MAE) in developing vision tasks such as image and video, exploring MAE in large-scale 3D point clouds remains challenging due to the inherent irregularity. In contrast to previous 3D MAE frameworks, which either design a complex decoder to infer masked information from maintained regions or adopt sophisticated masking strategies, we instead propose a much simpler paradigm. The core idea is to apply a \textbf{G}enerative \textbf{D}ecoder for MAE (GD-MAE) to automatically merges the surrounding context to restore the masked geometric knowledge in a hierarchical fusion manner. In doing so, our approach is free from introducing the heuristic design of decoders and enjoys the flexibility of exploring various masking strategies. The corresponding part costs less than \textbf{12\%} latency compared with conventional methods, while achieving better performance. We demonstrate the efficacy of the proposed method on several large-scale benchmarks: Waymo, KITTI, and ONCE. Consistent improvement on downstream detection tasks illustrates strong robustness and generalization capability. Not only our method reveals state-of-the-art results, but remarkably, we achieve comparable accuracy even with \textbf{20\%} of the labeled data on the Waymo dataset.
[](https://arxiv.org/abs/2212.03010) [](https://github.com/Nightmare-n/GD-MAE)
- **ALSO: Automotive Lidar Self-supervision by Occupancy estimation**.
Abstract
We propose a new self-supervised method for pre-training the backbone of deep perception models operating on point clouds. The core idea is to train the model on a pretext task which is the reconstruction of the surface on which the 3D points are sampled, and to use the underlying latent vectors as input to the perception head. The intuition is that if the network is able to reconstruct the scene surface, given only sparse input points, then it probably also captures some fragments of semantic information, that can be used to boost an actual perception task. This principle has a very simple formulation, which makes it both easy to implement and widely applicable to a large range of 3D sensors and deep networks performing semantic segmentation or object detection. In fact, it supports a single-stream pipeline, as opposed to most contrastive learning approaches, allowing training on limited resources. We conducted extensive experiments on various autonomous driving datasets, involving very different kinds of lidars, for both semantic segmentation and object detection. The results show the effectiveness of our method to learn useful representations without any annotation, compared to existing approaches.
