Finer-CAM: Spotting the Difference Reveals Finer Details for Visual Explanation

2025-01-20

AI Summary

Existing Class Activation Mapping (CAM) methods struggle to precisely localize subtle inter-class distinctions in fine-grained classification, resulting in insufficient interpretability. To address this, we propose a “difference-driven explanation” framework that, for the first time, integrates cross-class feature differencing into the CAM paradigm—without modifying the backbone network or requiring auxiliary supervision. Our approach explicitly suppresses shared responses between the target class and visually similar classes, thereby enhancing visualization of class-discriminative regions unique to the target. The method is compatible with mainstream CAM variants (e.g., Grad-CAM), supports adaptive weight scaling and tunable contrast intensity, and jointly improves both coarse contour and fine-grained spatial localization. On fine-grained benchmarks, masking the top 5% most activated pixels reduces the target-class confidence by 32% more on average than baseline methods, significantly strengthening causal interpretability and spatial localization accuracy.

Technology Category

• Computer Vision: Interpretability, Explainability, and Transparency
• Machine Learning: Transparent, Interpretable, Explainable ML
• Humans and AI: Explainable AI (XAI) for Human Understanding

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods

Abstract

Class activation map (CAM) has been widely used to highlight image regions that contribute to class predictions. Despite its simplicity and computational efficiency, CAM often struggles to identify discriminative regions that distinguish visually similar fine-grained classes. Prior efforts address this limitation by introducing more sophisticated explanation processes, but at the cost of extra complexity. In this paper, we propose Finer-CAM, a method that retains CAM's efficiency while achieving precise localization of discriminative regions. Our key insight is that the deficiency of CAM lies not in"how"it explains, but in"what"it explains}. Specifically, previous methods attempt to identify all cues contributing to the target class's logit value, which inadvertently also activates regions predictive of visually similar classes. By explicitly comparing the target class with similar classes and spotting their differences, Finer-CAM suppresses features shared with other classes and emphasizes the unique, discriminative details of the target class. Finer-CAM is easy to implement, compatible with various CAM methods, and can be extended to multi-modal models for accurate localization of specific concepts. Additionally, Finer-CAM allows adjustable comparison strength, enabling users to selectively highlight coarse object contours or fine discriminative details. Quantitatively, we show that masking out the top 5% of activated pixels by Finer-CAM results in a larger relative confidence drop compared to baselines. The source code and demo are available at https://github.com/Imageomics/Finer-CAM.

Problem

• Category Activation Map
• Similar Class Distinction
• Visual Interpretation Accuracy

Innovation

• Finer-CAM
• Category Discrimination
• Contrast Adjustment