When Epipolar Constraint Meets Non-local Operators in Multi-View Stereo

TL;DR

This paper introduces an epipolar line-constrained non-local feature augmentation method for multi-view stereo, significantly improving efficiency and accuracy by leveraging epipolar geometry to limit the search space.

Contribution

The paper proposes a novel epipolar line-based non-local augmentation strategy and an Epipolar Transformer, enhancing multi-view stereo performance while reducing computational overhead.

Findings

Achieves state-of-the-art results on DTU and Tanks-and-Temples benchmarks.

Reduces computational complexity by constraining non-local operations to epipolar lines.

Demonstrates improved reconstruction accuracy with efficient feature aggregation.

Abstract

Learning-based multi-view stereo (MVS) method heavily relies on feature matching, which requires distinctive and descriptive representations. An effective solution is to apply non-local feature aggregation, e.g., Transformer. Albeit useful, these techniques introduce heavy computation overheads for MVS. Each pixel densely attends to the whole image. In contrast, we propose to constrain non-local feature augmentation within a pair of lines: each point only attends the corresponding pair of epipolar lines. Our idea takes inspiration from the classic epipolar geometry, which shows that one point with different depth hypotheses will be projected to the epipolar line on the other view. This constraint reduces the 2D search space into the epipolar line in stereo matching. Similarly, this suggests that the matching of MVS is to distinguish a series of points lying on the same line. Inspired by this point-to-line search, we devise a line-to-point non-local augmentation strategy. We first devise an optimized searching algorithm to split the 2D feature maps into epipolar line pairs. Then, an Epipolar Transformer (ET) performs non-local feature augmentation among epipolar line pairs. We incorporate the ET into a learning-based MVS baseline, named ET-MVSNet. ET-MVSNet achieves state-of-the-art reconstruction performance on both the DTU and Tanks-and-Temples benchmark with high efficiency. Code is available at https://github.com/TQTQliu/ET-MVSNet.