MVImgNet2.0: A Larger-scale Dataset of Multi-view Images

Project Page | Paper (Arxiv)

by Xiaoguang Han*#, Yushuang Wu*, Luyue Shi*, Haolin Liu*, Hongjie Liao, Lingteng Qiu, Weihao Yuan, Xiaodong Gu, Zilong Dong, Shuguang Cui from GAP-Lab.

Introduction

MVImgNet2.0 contains ∼300k real-world objects in 340+ classes, expands MVImgNet to a total of ~520k real-life objects and 515 categories. The annotation comprehensively covers object masks, camera parameters, and point clouds.

Please fill out this form to get the download link and password.

Updates

• We have uploaded the script for downloading MVImgNet2.0.
• We have released the first part of MVImgNet2.0, which contains ~180k videos. Please fill out the above form to get the download links. We will release the remaining videos recently.

Folder structure

|-- ROOT
    |-- class_label
        |-- instance_id
            |-- images
            |-- masks
            |-- sparse/0 # camera parameters and sparse point clouds in colmap format
                |-- cameras.bin   
                |-- images.bin    
                |-- points3D.bin   

The mapping between class_label and class name can be found in mvimgnet2_category.json (The json file contains all the categories of MVImgNet and MVImgNet2.0, to use the data of MVImgNet, please refer to MVImgNet for dataset download).

The images folder contains the multi-view images, the masks folder contains the corresponding object masks, and the sparse folder contains the reconstructed camera parameters in COLMAP format. It is recommended to use the functions provided by COLMAP to read the binary files under sparse folder.

License

The data is released under the MVImgNet2.0 Terms of Use, and the code is released under the Attribution-NonCommercial 4.0 International License.

Copyright (c) 2024