Regional Multi-scale Approach for Visually Pleasing Explanations of Deep Neural Networks

TL;DR

This paper introduces a regional multi-scale approach that combines multi-scale segmentation and a prediction difference method to generate more class-discriminative and visually appealing explanations for deep neural network predictions.

Contribution

It proposes a novel regional multi-scale method that enhances explanation quality by fusing saliency maps from various segmentations, improving interpretability and visual appeal.

Findings

Produces more class-discriminative saliency maps

Generates visually pleasing explanations

Effective across different neural network models

Abstract

Recently, many methods to interpret and visualize deep neural network predictions have been proposed and significant progress has been made. However, a more class-discriminative and visually pleasing explanation is required. Thus, this paper proposes a region-based approach that estimates feature importance in terms of appropriately segmented regions. By fusing the saliency maps generated from multi-scale segmentations, a more class-discriminative and visually pleasing map is obtained. We incorporate this regional multi-scale concept into a prediction difference method that is model-agnostic. An input image is segmented in several scales using the super-pixel method, and exclusion of a region is simulated by sampling a normal distribution constructed using the boundary prior. The experimental results demonstrate that the regional multi-scale method produces much more class-discriminative and visually pleasing saliency maps.