# DepthCut: Improved Depth Edge Estimation Using Multiple Unreliable Channels

## TL;DR

DepthCut introduces a neural network-based method to accurately extract depth edges from unreliable channels, enhancing scene segmentation and depth estimation near discontinuities.

## Contribution

It presents a novel data-driven fusion approach for high-precision depth edges using multiple unreliable channels, improving downstream scene understanding tasks.

## Key findings

- Improved segmentation performance over 15 baseline variants.
- Enhanced depth estimation accuracy near depth edges.
- Qualitative results show superior scene segmentation and depth ordering.

## Abstract

In the context of scene understanding, a variety of methods exists to estimate different information channels from mono or stereo images, including disparity, depth, and normals. Although several advances have been reported in the recent years for these tasks, the estimated information is often imprecise particularly near depth discontinuities or creases. Studies have however shown that precisely such depth edges carry critical cues for the perception of shape, and play important roles in tasks like depth-based segmentation or foreground selection. Unfortunately, the currently extracted channels often carry conflicting signals, making it difficult for subsequent applications to effectively use them. In this paper, we focus on the problem of obtaining high-precision depth edges (i.e., depth contours and creases) by jointly analyzing such unreliable information channels. We propose DepthCut, a data-driven fusion of the channels using a convolutional neural network trained on a large dataset with known depth. The resulting depth edges can be used for segmentation, decomposing a scene into depth layers with relatively flat depth, or improving the accuracy of the depth estimate near depth edges by constraining its gradients to agree with these edges. Quantitatively, we compare against 15 variants of baselines and demonstrate that our depth edges result in an improved segmentation performance and an improved depth estimate near depth edges compared to data-agnostic channel fusion. Qualitatively, we demonstrate that the depth edges result in superior segmentation and depth orderings.