https://github.com/turingmotors/NuScenes-MQA

Official repository for the NuScenes-MQA. This paper is accepted by LLVA-AD Workshop at WACV 2024.
https://github.com/turingmotors/NuScenes-MQA

Last synced: 2 months ago
JSON representation

Official repository for the NuScenes-MQA. This paper is accepted by LLVA-AD Workshop at WACV 2024.

• Host: GitHub
• URL: https://github.com/turingmotors/NuScenes-MQA
• Owner: turingmotors
• License: apache-2.0
• Created: 2023-10-23T17:11:07.000Z (over 1 year ago)
• Default Branch: main
• Last Pushed: 2023-12-21T00:55:32.000Z (about 1 year ago)
• Last Synced: 2024-08-04T09:01:05.951Z (6 months ago)
• Homepage:
• Size: 49.8 KB
• Stars: 18
• Watchers: 1
• Forks: 0
• Open Issues: 2
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

• awesome-knowledge-driven-AD - NuScenes-MQA - 4, METEOR, ROUGE| (:books: Papers / Dataset \& Benchmark)

README

        
# NuScenes-MQA

![Sample Annotations](./src/sample_annotations.jpg)

## Abstract

Visual Question Answering (VQA) is one of the most important tasks in autonomous driving, which requires accurate recognition and complex situation evaluations. However, datasets annotated in a QA format, which guarantees precise language generation and scene recognition from driving scenes, have not been established yet. In this work, we introduce Markup-QA, a novel dataset annotation technique in which QAs are enclosed within markups. This approach facilitates the simultaneous evaluation of a model's capabilities in sentence generation and VQA. Moreover, using this annotation methodology, we designed the NuScenes-MQA dataset. This dataset empowers the development of vision language models, especially for autonomous driving tasks, by focusing on both descriptive capabilities and precise QA.

## Markup-QA Annotation

NuScenes-MQA annotations are available from [here](https://drive.google.com/drive/u/0/folders/1PQy0qhTtbdueIVlVnn4jC6xvANZUynRZ).

## Paper information

- **NuScenes-MQA: Integrated Evaluation of Captions and QA for Autonomous Driving Datasets using Markup Annotations**

- **arXiv**: [https://arxiv.org/abs/2312.06352](https://arxiv.org/abs/2312.06352)

- **Slides**: [link](https://docs.google.com/presentation/d/1mUtU9S7VVBmDy7nZo_PzmuVxSpeQracD9x_1B2zc-2g/edit#slide=id.g2a7ef2f5709_0_226)

## This paper was accepted at the LLVM-AD Workshop at WACV

- **Workshop**: LLVM-AD at WACV

- **Year**: 2024/1

- **Website**: [https://llvm-ad.github.io/](https://llvm-ad.github.io/)

## BibTeX

```bibtex

@InProceedings{Inoue_2024_WACV,

    author    = {Inoue, Yuichi and Yada, Yuki and Tanahashi, Kotaro and Yamaguchi, Yu},

    title     = {NuScenes-MQA: Integrated Evaluation of Captions and QA for Autonomous Driving Datasets Using Markup Annotations},

    booktitle = {Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) Workshops},

    month     = {January},

    year      = {2024},

    pages     = {930-938}

}

```