Axiom-based Grad-CAM: Towards Accurate Visualization and Explanation of CNNs

TL;DR

This paper introduces XGrad-CAM, an improved visualization method for CNNs that incorporates axioms to enhance accuracy, class-discrimination, and theoretical grounding, outperforming existing CAM-based methods.

Contribution

It proposes axioms for CNN visualization and develops XGrad-CAM, a method that better satisfies these axioms and improves visualization quality over Grad-CAM and related techniques.

Findings

XGrad-CAM outperforms Grad-CAM in visualization accuracy.

XGrad-CAM is class-discriminative and easy to implement.

Theoretical axioms improve CNN visualization methods.

Abstract

To have a better understanding and usage of Convolution Neural Networks (CNNs), the visualization and interpretation of CNNs has attracted increasing attention in recent years. In particular, several Class Activation Mapping (CAM) methods have been proposed to discover the connection between CNN's decision and image regions. In spite of the reasonable visualization, lack of clear and sufficient theoretical support is the main limitation of these methods. In this paper, we introduce two axioms -- Conservation and Sensitivity -- to the visualization paradigm of the CAM methods. Meanwhile, a dedicated Axiom-based Grad-CAM (XGrad-CAM) is proposed to satisfy these axioms as much as possible. Experiments demonstrate that XGrad-CAM is an enhanced version of Grad-CAM in terms of conservation and sensitivity. It is able to achieve better visualization performance than Grad-CAM, while also be class-discriminative and easy-to-implement compared with Grad-CAM++ and Ablation-CAM. The code is available at https://github.com/Fu0511/XGrad-CAM.