Dancing-Huggingface/EgoDynamic4D

1---2task_categories:3- question-answering4language:5- en6---7# EgoDynamic4D Dataset from AAAI 2026 paper: [Understanding Dynamic Scenes in Egocentric 4D Point Clouds](https://arxiv.org/abs/2508.07251)8 9🚀 **EgoDynamic4D QA dataset has been released.**10 11This repository hosts the official implementation of **EgoDynamic4D**, a large-scale egocentric **4D dynamic scene understanding** benchmark introduced in our AAAI 2026 paper:12 13> **Understanding Dynamic Scenes in Egocentric 4D Point Clouds**14 15---16 17## About the Dataset18 19**EgoDynamic4D** is a question answering (QA) benchmark designed for fine-grained **spatio-temporal reasoning** in egocentric dynamic scenes.20 21The dataset includes:22 23* Egocentric RGB-D videos24* Camera poses25* Globally unique instance masks26* 4D bounding boxes over time27* Large-scale **QA annotations for dynamic reasoning** tasks28 29The dataset is constructed based on existing egocentric 4D resources, building upon [ADT](https://arxiv.org/abs/2306.06362) and [THUD++](https://arxiv.org/abs/2412.08096). Our main contribution lies in the **large-scale, task-driven QA annotations**, which enable fine-grained spatio-temporal reasoning in dynamic egocentric scenes.30 31**Please find instructions on [Github](https://github.com/Dancing-Github/EgoDynamic4D) for usages.**