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

TL;DR

MVImgNet2.0 is a significantly expanded and improved multi-view image dataset with 520k objects, supporting advanced 3D reconstruction and vision tasks through high-quality multi-view captures, segmentation, and pose estimation.

Contribution

The paper introduces MVImgNet2.0, a larger, higher-quality multi-view dataset with 360-degree views, improved segmentation, and better pose estimation, bridging 2D and 3D vision research.

Findings

Enhances 3D reconstruction performance with the new dataset.

Provides high-quality multi-view images and point clouds for research.

Supports diverse downstream 3D vision applications.

Abstract

MVImgNet is a large-scale dataset that contains multi-view images of ~220k real-world objects in 238 classes. As a counterpart of ImageNet, it introduces 3D visual signals via multi-view shooting, making a soft bridge between 2D and 3D vision. This paper constructs the MVImgNet2.0 dataset that expands MVImgNet into a total of ~520k objects and 515 categories, which derives a 3D dataset with a larger scale that is more comparable to ones in the 2D domain. In addition to the expanded dataset scale and category range, MVImgNet2.0 is of a higher quality than MVImgNet owing to four new features: (i) most shoots capture 360-degree views of the objects, which can support the learning of object reconstruction with completeness; (ii) the segmentation manner is advanced to produce foreground object masks of higher accuracy; (iii) a more powerful structure-from-motion method is adopted to derive the camera pose for each frame of a lower estimation error; (iv) higher-quality dense point clouds are reconstructed via advanced methods for objects captured in 360-degree views, which can serve for downstream applications. Extensive experiments confirm the value of the proposed MVImgNet2.0 in boosting the performance of large 3D reconstruction models. MVImgNet2.0 will be public at luyues.github.io/mvimgnet2, including multi-view images of all 520k objects, the reconstructed high-quality point clouds, and data annotation codes, hoping to inspire the broader vision community.