Towards Interpretable Deep Networks for Monocular Depth Estimation

TL;DR

This paper introduces a method to enhance the interpretability of deep networks for monocular depth estimation by assigning depth ranges to hidden units, improving understanding without sacrificing accuracy.

Contribution

It proposes a novel approach to train interpretable depth estimation networks by leveraging depth selectivity of hidden units without altering the original architecture.

Findings

Improved depth selectivity of hidden units.

Enhanced interpretability without loss of accuracy.

Method applicable across different models and datasets.

Abstract

Deep networks for Monocular Depth Estimation (MDE) have achieved promising performance recently and it is of great importance to further understand the interpretability of these networks. Existing methods attempt to provide posthoc explanations by investigating visual cues, which may not explore the internal representations learned by deep networks. In this paper, we find that some hidden units of the network are selective to certain ranges of depth, and thus such behavior can be served as a way to interpret the internal representations. Based on our observations, we quantify the interpretability of a deep MDE network by the depth selectivity of its hidden units. Moreover, we then propose a method to train interpretable MDE deep networks without changing their original architectures, by assigning a depth range for each unit to select. Experimental results demonstrate that our method is able to enhance the interpretability of deep MDE networks by largely improving the depth selectivity of their units, while not harming or even improving the depth estimation accuracy. We further provide a comprehensive analysis to show the reliability of selective units, the applicability of our method on different layers, models, and datasets, and a demonstration on analysis of model error. Source code and models are available at https://github.com/youzunzhi/InterpretableMDE .