Update README.md

README.md

@@ -1,7 +1,7 @@
 ---
 task_categories:
 - object-detection
-license: unknown
+license: cc-by-4.0
 tags:
 - 3d-object-detection
 - computer-vision
@@ -13,76 +13,167 @@ language:
 - en
 ---
 
-# M2I Dataset
+---
+
+# M2I: Multi-camera Multi-layout Infrastructure Dataset
 
-The M2I dataset is a synthetic dataset introduced in the paper [MIC-BEV: Multi-Infrastructure Camera Bird's-Eye-View Transformer with Relation-Aware Fusion for 3D Object Detection](https://huggingface.co/papers/2510.24688).
+[![Paper](https://img.shields.io/badge/Paper-arXiv-b31b1b.svg)](https://arxiv.org/abs/2510.24688)
+[![Code](https://img.shields.io/badge/GitHub-MIC--BEV-blue.svg)](https://github.com/HandsomeYun/MIC-BEV)
 
-M2I is designed to support training and evaluation of models for infrastructure-based multi-camera 3D object detection. It features diverse camera configurations, road layouts, and environmental conditions. This dataset is crucial for developing robust perception systems in intelligent transportation environments and is used in conjunction with frameworks like MIC-BEV.
+The **M2I (Multi-camera Multi-layout Infrastructure)** dataset is a large-scale synthetic dataset for **infrastructure-based multi-camera 3D object detection**. It is introduced in our paper:
 
-## Paper
+> **MIC-BEV: Multi-Infrastructure Camera Bird's-Eye-View Transformer with Relation-Aware Fusion for 3D Object Detection**
 
-[MIC-BEV: Multi-Infrastructure Camera Bird's-Eye-View Transformer with Relation-Aware Fusion for 3D Object Detection](https://huggingface.co/papers/2510.24688)
+M2I is designed to support research on intelligent transportation systems by providing diverse infrastructure camera deployments, road layouts, environmental conditions, and traffic scenarios for Bird's-Eye-View (BEV) perception.
+---
+# Overview
 
-## Code
+M2I provides synchronized multi-camera observations together with 3D object annotations for infrastructure perception.
 
-The official code repository for MIC-BEV, which utilizes this dataset, can be found here: [https://github.com/HandsomeYun/MIC-BEV](https://github.com/HandsomeYun/MIC-BEV)
+The dataset includes
 
-## Sample Usage
+- Multi-camera RGB images
+- Camera calibration (intrinsics and extrinsics)
+- 3D bounding box annotations
+- BEV semantic maps
+- Temporal sequences
+- Camera visibility information
+---
+# Dataset Features
 
-The M2I dataset is intended to be used with the MIC-BEV codebase for training and evaluation of 3D object detection models. The following snippets demonstrate how to prepare the dataset and perform quick training/evaluation, as outlined in the [MIC-BEV GitHub repository](https://github.com/HandsomeYun/MIC-BEV).
+Compared with existing infrastructure perception datasets, M2I provides:
+
+- Diverse road layouts (intersections, highways, roundabouts, ramps, etc.)
+- Variable camera configurations (1–4 cameras)
+- Heterogeneous camera intrinsics and extrinsics
+- Multiple camera mounting locations and heights
+- Diverse weather conditions
+- Diverse illumination conditions
+- Dynamic traffic density
+- Sensor degradation (blur, masking, calibration perturbation)
+
+---
 
-### Prepare Dataset
+# Dataset Structure
+
+After downloading the dataset from Hugging Face, the repository contains:
+
+```text
+M2I/
+
+├── M2I_split_dataset.tar.gz.part-*
+├── README.md
+```
+
+The RGB images are stored as multipart archives to simplify distribution on Hugging Face.
+
+After extracting the archive, the dataset is organized as
+
+```text
+M2I/
+
+├── M2I_split_dataset/
+│   ├── train/
+│   ├── val/
+│   └── test/
+├── M2I_json/
+├── Town01.xodr
+├── Town03.xodr
+├── Town04.xodr
+├── Town05.xodr
+├── Town06.xodr
+├── Town07.xodr
+├── Town10HD_Opt.xodr
+└── README.md
+```
 
-To prepare the M2I-style data (e.g., V2XSet), use the provided preprocessing script from the MIC-BEV repository:
+where
+
+- **M2I_split_dataset/** contains the RGB images and annotations.
+- **M2I_json/** contains the raw frame-level metadata before conversion to training annotations.
+- **Town0X.xodr** and **Town*.xodr** contain the CARLA HD maps used to generate the scenes.
+
+---
+
+# Download
+
+Download the dataset using the Hugging Face CLI:
 
 ```bash
-python MIC-BEV_Official/tools/data_converter/mic-bev/create_v2xset_multiple_map.py
+hf download handsomeYun/M2I \
+  --repo-type dataset \
+  --local-dir /path/to/M2I
 ```
-This script processes raw V2XSet-format data and generates the necessary multi-map annotations for training MIC-BEV with M2I.
-
-### Quick Start (Training and Evaluation)
-
-After setting up the environment (refer to the GitHub repository's `Installation` section for full details) and preparing the dataset, you can use the MIC-BEV configuration files for training and evaluation:
-
-1.  **Training**: Use the MIC-BEV configuration file:
-    ```bash
-    python tools/train.py projects/configs/micbev/mic-bev-seg-gnn.py
-    ```
-
-2.  **Evaluation**: Evaluate your trained model:
-    ```bash
-    python tools/test.py projects/configs/micbev/mic-bev-seg-gnn.py /path/to/checkpoint.pth --eval bbox
-    ```
-
-
-## Citation
- If you are using our OpenCDA framework or codes for your development, please cite the following paper:
- ```bibtex
-@misc{zhang2025micbev,
-  title        = {MIC-BEV: Multi-Infrastructure Camera Bird's-Eye-View Transformer with Relation-Aware Fusion for 3D Object Detection},
-  author       = {Zhang, Yun and Zheng, Zhaoliang and Liu, Johnson and Huang, Zhiyu and Zhou, Zewei and Meng, Zonglin and Cai, Tianhui and Ma, Jiaqi},
-  year         = {2025},
-  eprint       = {2510.24688},
-  archivePrefix= {arXiv},
-  primaryClass = {cs.CV},
-  url          = {https://arxiv.org/abs/2510.24688}
-}
 
+Extract the image archive:
+
+```bash
+mkdir -p /path/to/M2I_data
+
+cat M2I_split_dataset.tar.gz.part-* | tar -xzvf - -C /path/to/M2I_data
 ```
-The arxiv link to the paper:  [https://arxiv.org/abs/2107.06260](https://arxiv.org/abs/2510.24688)
 
-## License
+---
 
-This project is released under a UCLA research license.
+# Preparing Training Annotations
 
-The MIC-BEV codebase is provided for non-commercial academic research only.
-Researchers are permitted to use and modify the source code for their own research purposes.
+After extracting the dataset, generate the training annotations using the MIC-BEV preprocessing script:
 
-The following uses are not permitted under this license:
-- Commercial use of any kind
-- Contracted work that generates corporate or institutional revenue
-- Redistribution for commercial products or services
-- Please refer to the LICENSE file for full terms and for information on potential commercial licensing opportunities.
+```bash
+python tools/data_converter/mic-bev/create_v2xset_multiple_map.py
+```
 
-## Contributors
-MIC-BEV is mainly supported by the [UCLA Mobility Lab](https://mobility-lab.seas.ucla.edu/). <br>
+Please refer to the MIC-BEV repository for detailed training instructions.
+
+---
+
+# Related Project
+
+The official implementation of MIC-BEV is available at
+
+https://github.com/HandsomeYun/MIC-BEV
+
+The repository contains
+
+- Installation instructions
+- Dataset preprocessing
+- Training
+- Evaluation
+- Visualization
+
+---
+
+# License
+
+The M2I dataset is released under the **CC BY 4.0** license.
+
+Please also comply with the licenses of any third-party software or datasets used together with this project.
+
+---
+
+# Acknowledgements
+
+The M2I dataset was developed by the UCLA Mobility Lab as part of the MIC-BEV project.
+
+MIC-BEV builds upon several outstanding open-source projects, including:
+
+- BEVFormer
+- DETR3D
+- DD3D
+- MMDetection3D
+- RoScenes
+
+---
+
+# Citation
+
+If you use M2I in your research, please cite:
+
+```bibtex
+@article{zhang2025micbev,
+  title={MIC-BEV: Multi-Infrastructure Camera Bird's-Eye-View Transformer with Relation-Aware Fusion for 3D Object Detection},
+  author={Zhang, Yun and Zheng, Zhaoliang and Liu, Johnson and Huang, Zhiyu and Zhou, Zewei and Meng, Zonglin and Cai, Tianhui and Ma, Jiaqi},
+  journal={arXiv preprint arXiv:2510.24688},
+  year={2025}
+}
+```