Learning sparse representations of depth

TL;DR

This paper presents a novel method for learning sparse depth representations that handle spatially varying noise, improving depth map denoising and inference when integrated with existing stereo matching algorithms.

Contribution

It introduces a non-stationary sparse coding approach for depth maps, enhancing depth inference by capturing higher-order dependencies and integrating with MRF-based methods.

Findings

Achieves state-of-the-art depth map denoising results.

Improves depth estimation accuracy when combined with existing algorithms.

Handles spatially varying noise in depth data effectively.

Abstract

This paper introduces a new method for learning and inferring sparse representations of depth (disparity) maps. The proposed algorithm relaxes the usual assumption of the stationary noise model in sparse coding. This enables learning from data corrupted with spatially varying noise or uncertainty, typically obtained by laser range scanners or structured light depth cameras. Sparse representations are learned from the Middlebury database disparity maps and then exploited in a two-layer graphical model for inferring depth from stereo, by including a sparsity prior on the learned features. Since they capture higher-order dependencies in the depth structure, these priors can complement smoothness priors commonly used in depth inference based on Markov Random Field (MRF) models. Inference on the proposed graph is achieved using an alternating iterative optimization technique, where the first layer is solved using an existing MRF-based stereo matching algorithm, then held fixed as the second layer is solved using the proposed non-stationary sparse coding algorithm. This leads to a general method for improving solutions of state of the art MRF-based depth estimation algorithms. Our experimental results first show that depth inference using learned representations leads to state of the art denoising of depth maps obtained from laser range scanners and a time of flight camera. Furthermore, we show that adding sparse priors improves the results of two depth estimation methods: the classical graph cut algorithm by Boykov et al. and the more recent algorithm of Woodford et al.