Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/fudan-zvg/Ego3RT
[ECCV 2022] Learning Ego 3D Representation as Ray Tracing
https://github.com/fudan-zvg/Ego3RT
Last synced: about 2 months ago
JSON representation
[ECCV 2022] Learning Ego 3D Representation as Ray Tracing
- Host: GitHub
- URL: https://github.com/fudan-zvg/Ego3RT
- Owner: fudan-zvg
- License: mit
- Created: 2022-06-06T12:24:23.000Z (over 2 years ago)
- Default Branch: main
- Last Pushed: 2022-09-15T15:37:04.000Z (about 2 years ago)
- Last Synced: 2024-08-01T03:42:41.329Z (4 months ago)
- Language: Python
- Homepage:
- Size: 27.3 MB
- Stars: 105
- Watchers: 12
- Forks: 7
- Open Issues: 3
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
- Awesome-BEV-Perception - project
README
# Learning Ego 3D Representation as Ray Tracing
### [Website](https://fudan-zvg.github.io/Ego3RT) | [Paper](https://arxiv.org/abs/2206.04042)
> [**Learning Ego 3D Representation as Ray Tracing**](https://arxiv.org/abs/2206.04042),
> Jiachen Lu, Zheyuan Zhou, Xiatian Zhu, Hang Xu, Li Zhang
> **ECCV 2022**
## Demo
## Video
[](https://www.youtube.com/watch?v=lMwcGUahlJg)
## News
- [2022/07/19]: Configs and instructions for training are released!
- [2022/07/05]: First version of Ego3RT is released! Code for detection head and training configs will comming soon.
- [2022/07/04]: Ego3RT is accepted by **ECCV 2022**!
## Abstract
A self-driving perception model aims to extract 3D semantic representations from multiple cameras collectively into the bird's-eye-view (BEV) coordinate frame of the ego car in order to ground downstream planner. Existing perception methods often rely on error-prone depth estimation of the whole scene or learning sparse virtual 3D representations without the target geometry structure, both of which remain limited in performance and/or capability. In this paper, we present a novel end-to-end architecture for ego 3D representation learning from an arbitrary number of unconstrained camera views. Inspired by the ray tracing principle, we design a polarized grid of ``imaginary eyes" as the learnable ego 3D representation and formulate the learning process with the adaptive attention mechanism in conjunction with the 3D-to-2D projection. Critically, this formulation allows extracting rich 3D representation from 2D images without any depth supervision, and with the built-in geometry structure consistent w.r.t. BEV. Despite its simplicity and versatility, extensive experiments on standard BEV visual tasks (e.g., camera-based 3D object detection and BEV segmentation) show that our model outperforms all state-of-the-art alternatives significantly, with an extra advantage in computational efficiency from multi-task learning.
## Methods
## Train & Test
Please refer to the [get_started.md](get_started.md)
## Result
### **3D object detection on nuScenes validation set**
| Model | Polar size | mAP | NDS | checkpoint |
| :--------------------------------------------------------------------------------: | :--------: | :--: | :--: | :----------------------------------------------------------------------------------------------------------------------: |
| [Ego3RT, ResNet101_DCN](projects/configs/ego3rt/ego3rt_polar80x256_cart160x160.py) | 80x256 | 37.5 | 45.0 | |
| [Ego3RT, ResNet101_DCN](projects/configs/ego3rt/ego3rt_polar72x192_cart128x128.py) | 72x192 | 37.5 | 44.9 | [ego3rt_polar72x192_cart128x128.pth](https://drive.google.com/file/d/1KFApMbHCDUdesT81wDUunJB09yo01FPW/view?usp=sharing) |
| Ego3RT, VoVNet | 80x256 | 47.8 | 53.4 | |
### **3D object detection on nuScenes test set**
| Model | Polar size | mAP | NDS |
| :--------------------------------------------------------------------------------: | :--------: | :--: | ---- |
| [Ego3RT, ResNet101_DCN](projects/configs/ego3rt/ego3rt_polar80x256_cart160x160.py) | 80x256 | 38.9 | 44.3 |
| Ego3RT, VoVNet | 80x256 | 42.5 | 47.3 |
### **BEV segmentation on nuScenes validation set**
| Model | Polar size | Multitask | mIoU |
| :-------------------: | :--------: | :-------: | :--: |
| Ego3RT, EfficientNet | 80x256 | no | 55.5 |
| Ego3RT, ResNet101_DCN | 80x256 | yes | 46.2 |
## License
[MIT](LICENSE)
## Reference
```bibtex
@inproceedings{lu2022ego3rt,
title={Learning Ego 3D Representation as Ray Tracing},
author={Lu, Jiachen and Zhou, Zheyuan and Zhu, Xiatian and Xu, Hang and Zhang, Li},
booktitle={European Conference on Computer Vision},
year={2022}
}
```
## Acknowledgement
Thanks to previous open-sourced repo:
* [MMDetection3D](https://github.com/open-mmlab/mmdetection3d)
* [DETR3D](https://github.com/WangYueFt/detr3d)
* [Deformable DETR](https://github.com/fundamentalvision/Deformable-DETR)