Regional multi-scale approach for visually pleasing explanations of deep neural networks

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. This paper incorporates this regional multi-scale concept into a prediction difference method that is model-agnostic. An input image is segmented in several scales using the superpixel method, and exclusion of a region is simulated by sampling a normal distribution constructed via the boundary prior. The experimental results demonstrate that the regional multi-scale method produces much more class-discriminative and visually pleasing saliency maps.