Learning Unsupervised Multi-View Stereopsis via Robust Photometric Consistency

TL;DR

This paper introduces an unsupervised learning method for multi-view stereopsis that uses robust photometric consistency to predict depth without requiring ground-truth 3D data, handling occlusions and lighting variations effectively.

Contribution

It proposes a novel robust loss formulation for unsupervised multi-view stereo learning that implicitly manages occlusions and lighting changes, enabling training without 3D supervision.

Findings

Achieves accurate depth prediction without ground-truth 3D data

Produces more complete reconstructions than ground-truth-based methods

Generalizes well to new datasets and allows unsupervised fine-tuning

Abstract

We present a learning based approach for multi-view stereopsis (MVS). While current deep MVS methods achieve impressive results, they crucially rely on ground-truth 3D training data, and acquisition of such precise 3D geometry for supervision is a major hurdle. Our framework instead leverages photometric consistency between multiple views as supervisory signal for learning depth prediction in a wide baseline MVS setup. However, naively applying photo consistency constraints is undesirable due to occlusion and lighting changes across views. To overcome this, we propose a robust loss formulation that: a) enforces first order consistency and b) for each point, selectively enforces consistency with some views, thus implicitly handling occlusions. We demonstrate our ability to learn MVS without 3D supervision using a real dataset, and show that each component of our proposed robust loss results in a significant improvement. We qualitatively observe that our reconstructions are often more complete than the acquired ground truth, further showing the merits of this approach. Lastly, our learned model generalizes to novel settings, and our approach allows adaptation of existing CNNs to datasets without ground-truth 3D by unsupervised finetuning. Project webpage: https://tejaskhot.github.io/unsup_mvs