CostFormer:Cost Transformer for Cost Aggregation in Multi-view Stereo

TL;DR

CostFormer introduces an efficient Transformer-based approach for cost aggregation in Multi-view Stereo, addressing CNN limitations and computational challenges to improve long-range feature aggregation.

Contribution

The paper proposes CostFormer, a novel Transformer-based cost aggregation network with RDACT and RRT modules, overcoming computational limits and enhancing MVS performance.

Findings

Effective long-range feature aggregation via self-attention.

Reduced memory usage and inference latency.

Universal plug-in for existing MVS methods.

Abstract

The core of Multi-view Stereo(MVS) is the matching process among reference and source pixels. Cost aggregation plays a significant role in this process, while previous methods focus on handling it via CNNs. This may inherit the natural limitation of CNNs that fail to discriminate repetitive or incorrect matches due to limited local receptive fields. To handle the issue, we aim to involve Transformer into cost aggregation. However, another problem may occur due to the quadratically growing computational complexity caused by Transformer, resulting in memory overflow and inference latency. In this paper, we overcome these limits with an efficient Transformer-based cost aggregation network, namely CostFormer. The Residual Depth-Aware Cost Transformer(RDACT) is proposed to aggregate long-range features on cost volume via self-attention mechanisms along the depth and spatial dimensions. Furthermore, Residual Regression Transformer(RRT) is proposed to enhance spatial attention. The proposed method is a universal plug-in to improve learning-based MVS methods.