Learning Monocular Depth by Distilling Cross-domain Stereo Networks

TL;DR

This paper introduces a novel framework that leverages stereo matching networks trained on synthetic data to supervise monocular depth estimation, effectively bridging domain gaps and achieving state-of-the-art results on KITTI.

Contribution

It proposes using stereo networks as proxies to transfer depth knowledge from synthetic to real data, enhancing monocular depth estimation accuracy.

Findings

Achieves state-of-the-art results on KITTI dataset.

Effectively utilizes synthetic data for monocular depth learning.

Demonstrates successful cross-domain knowledge transfer.

Abstract

Monocular depth estimation aims at estimating a pixelwise depth map for a single image, which has wide applications in scene understanding and autonomous driving. Existing supervised and unsupervised methods face great challenges. Supervised methods require large amounts of depth measurement data, which are generally difficult to obtain, while unsupervised methods are usually limited in estimation accuracy. Synthetic data generated by graphics engines provide a possible solution for collecting large amounts of depth data. However, the large domain gaps between synthetic and realistic data make directly training with them challenging. In this paper, we propose to use the stereo matching network as a proxy to learn depth from synthetic data and use predicted stereo disparity maps for supervising the monocular depth estimation network. Cross-domain synthetic data could be fully utilized in this novel framework. Different strategies are proposed to ensure learned depth perception capability well transferred across different domains. Our extensive experiments show state-of-the-art results of monocular depth estimation on KITTI dataset.