MVMO: A Multi-Object Dataset for Wide Baseline Multi-View Semantic Segmentation

TL;DR

MVMO is a large synthetic multi-view dataset designed to advance research in multi-view semantic segmentation by providing diverse, high-density scenes with wide camera baselines and complex occlusions.

Contribution

The paper introduces MVMO, a novel synthetic dataset with wide baselines and dense objects, enabling improved multi-view semantic segmentation research.

Findings

Baseline experiments show the need for new methods to leverage multi-view information.

High disparity and occlusion challenge single-view segmentation.

MVMO facilitates research in cross-view semantic transfer.

Abstract

We present MVMO (Multi-View, Multi-Object dataset): a synthetic dataset of 116,000 scenes containing randomly placed objects of 10 distinct classes and captured from 25 camera locations in the upper hemisphere. MVMO comprises photorealistic, path-traced image renders, together with semantic segmentation ground truth for every view. Unlike existing multi-view datasets, MVMO features wide baselines between cameras and high density of objects, which lead to large disparities, heavy occlusions and view-dependent object appearance. Single view semantic segmentation is hindered by self and inter-object occlusions that could benefit from additional viewpoints. Therefore, we expect that MVMO will propel research in multi-view semantic segmentation and cross-view semantic transfer. We also provide baselines that show that new research is needed in such fields to exploit the complementary information of multi-view setups.