# Learned Multi-Patch Similarity

## TL;DR

This paper introduces a learned multi-patch similarity function using neural networks for depth estimation from multiple views, outperforming traditional pairwise methods.

## Contribution

It proposes a novel neural network-based approach to directly compute similarity across multiple image patches, addressing a gap in multi-view depth estimation.

## Key findings

- Outperforms pairwise similarity methods on multi-view datasets
- Demonstrates advantages of learned multi-patch similarity
- Validates effectiveness across several datasets

## Abstract

Estimating a depth map from multiple views of a scene is a fundamental task in computer vision. As soon as more than two viewpoints are available, one faces the very basic question how to measure similarity across >2 image patches. Surprisingly, no direct solution exists, instead it is common to fall back to more or less robust averaging of two-view similarities. Encouraged by the success of machine learning, and in particular convolutional neural networks, we propose to learn a matching function which directly maps multiple image patches to a scalar similarity score. Experiments on several multi-view datasets demonstrate that this approach has advantages over methods based on pairwise patch similarity.