AcED: Accurate and Edge-consistent Monocular Depth Estimation

TL;DR

This paper introduces a fully differentiable ordinal regression approach for monocular depth estimation that produces smooth, edge-consistent depth maps in an end-to-end manner, outperforming previous methods.

Contribution

It presents the first fully differentiable ordinal regression framework for depth estimation, incorporating boundary and smoothness constraints for improved accuracy.

Findings

Outperforms recent state-of-the-art methods quantitatively and qualitatively.

Produces smooth, edge-consistent depth maps.

Demonstrates practical utility in single camera bokeh applications.

Abstract

Single image depth estimation is a challenging problem. The current state-of-the-art method formulates the problem as that of ordinal regression. However, the formulation is not fully differentiable and depth maps are not generated in an end-to-end fashion. The method uses a na\"ive threshold strategy to determine per-pixel depth labels, which results in significant discretization errors. For the first time, we formulate a fully differentiable ordinal regression and train the network in end-to-end fashion. This enables us to include boundary and smoothness constraints in the optimization function, leading to smooth and edge-consistent depth maps. A novel per-pixel confidence map computation for depth refinement is also proposed. Extensive evaluation of the proposed model on challenging benchmarks reveals its superiority over recent state-of-the-art methods, both quantitatively and qualitatively. Additionally, we demonstrate practical utility of the proposed method for single camera bokeh solution using in-house dataset of challenging real-life images.