Improve dataset card: add metadata and links to paper/project/code

README.md

@@ -1,15 +1,61 @@
 ---
 license: mit
+task_categories:
+- other
+tags:
+- computer-vision
+- point-tracking
+- event-camera
+- multimodal
 ---
-TAPFormer: Robust Arbitrary Point Tracking via Transient Asynchronous Fusion of Frames and Events (CVPR 2026)
 
-https://arxiv.org/abs/2603.04989
+# TAPFormer: Robust Arbitrary Point Tracking via Transient Asynchronous Fusion of Frames and Events
 
-Test datasets: InivTAP and DrivTAP
+[**Project Page**](https://tapformer.github.io/) | [**Paper**](https://huggingface.co/papers/2603.04989) | [**GitHub**](https://github.com/ljx1002/TAPFormer)
 
-Project website: tapformer.github.io
+TAPFormer is a transformer-based framework for robust and high-frequency arbitrary point tracking (TAP). It introduces a Transient Asynchronous Fusion (TAF) mechanism to bridge the gap between low-rate RGB frames and high-rate event streams.
 
-The github repo: https://github.com/ljx1002/TAPFormer
+This repository hosts the **InivTAP** and **DrivTAP** benchmarks, which are real-world frame-event TAP datasets covering diverse illumination and motion conditions.
 
+## Dataset Summary
+- **InivTAP**: Indoor sequences with various objects and lighting conditions, including synchronized frame-event data.
+- **DrivTAP**: Outdoor driving sequences captured under realistic conditions.
 
+## Dataset Structure
+To use these datasets with the TAPFormer codebase, ensure your data is organized as follows:
 
+```
+dataset_dir/
+├── InivTAP/
+│   └── sequence_name/
+│       ├── events/
+│       ├── images_corrected/
+│       └── annotations.npy
+└── DrivTAP/
+    └── sequence_name/
+        ├── events/
+        ├── images_corrected/
+        └── annotations.npy
+```
+
+## Citation
+
+If you use this dataset or code in your research, please cite:
+
+```bibtex
+@article{liu2026tapformer,
+  title={TAPFormer: Robust Arbitrary Point Tracking via Transient Asynchronous Fusion of Frames and Events},
+  author={Liu, Jiaxiong and Tan, Zhen and Zhang, Jinpu and Zhou, Yi and Shen, Hui bit and Chen, Xieyuanli and Hu, Dewen},
+  journal={arXiv preprint arXiv:2603.04989},
+  year={2026}
+}
+
+@inproceedings{liu2025tracking,
+  title={Tracking any point with frame-event fusion network at high frame rate},
+  author={Liu, Jiaxiong and Wang, Bo and Tan, Zhen and Zhang, Jinpu and Shen, Hui and Hu, Dewen},
+  booktitle={2025 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
+  pages={18834--18840},
+  year={2025},
+  organization={IEEE}
+}
+```