Skip to content

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

License

Notifications You must be signed in to change notification settings

turingmotors/NuScenes-MQA

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 
 
 
 
 

Repository files navigation

NuScenes-MQA

Sample Annotations

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.

Paper information

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

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}
}

About

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

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published