How Good MVSNets Are at Depth Fusion

TL;DR

This paper investigates how incorporating low-quality sensor depth as additional input affects the performance of deep multi-view stereo methods, demonstrating potential improvements in depth estimation quality.

Contribution

The authors adapt two leading deep multi-view stereo methods to include sensor depth input, showing that this can enhance depth reconstruction accuracy.

Findings

Sensor depth input can improve stereo reconstruction quality.

Modifications to existing methods enable effective use of low-quality depth data.

Additional depth input may lead to more accurate 3D reconstructions.

Abstract

We study the effects of the additional input to deep multi-view stereo methods in the form of low-quality sensor depth. We modify two state-of-the-art deep multi-view stereo methods for using with the input depth. We show that the additional input depth may improve the quality of deep multi-view stereo.