Explanations for Occluded Images

TL;DR

This paper introduces a new black-box explanation algorithm based on causal theory, significantly improving interpretability of occluded images while maintaining competitive performance on clear images.

Contribution

The paper presents a novel causal-based explanation method for occluded images, implemented in the DEEPCOVER tool, outperforming existing explanation algorithms on occluded inputs.

Findings

More accurate explanations on occluded images

Performance comparable to state-of-the-art on unoccluded images

Effective in black-box settings

Abstract

Existing algorithms for explaining the output of image classifiers perform poorly on inputs where the object of interest is partially occluded. We present a novel, black-box algorithm for computing explanations that uses a principled approach based on causal theory. We have implemented the method in the DEEPCOVER tool. We obtain explanations that are much more accurate than those generated by the existing explanation tools on images with occlusions and observe a level of performance comparable to the state of the art when explaining images without occlusions.