Class Activation Map generation by Multiple Level Class Grouping and Orthogonal Constraint

TL;DR

This paper introduces a novel CAM generation method using multiple hierarchical classification models and an orthogonal constraint to improve the localization of class regions in weakly supervised tasks.

Contribution

It proposes a multi-model CAM generation framework based on class semantic relationships and an orthogonal module to enhance region localization.

Findings

Improved CAM quality on PASCAL VOC 2012 dataset

Enhanced object region extraction compared to baseline methods

Effective use of hierarchical class relationships for CAM generation

Abstract

Class activation map (CAM) highlights regions of classes based on classification network, which is widely used in weakly supervised tasks. However, it faces the problem that the class activation regions are usually small and local. Although several efforts paid to the second step (the CAM generation step) have partially enhanced the generation, we believe such problem is also caused by the first step (training step), because single classification model trained on the entire classes contains finite discriminate information that limits the object region extraction. To this end, this paper solves CAM generation by using multiple classification models. To form multiple classification networks that carry different discriminative information, we try to capture the semantic relationships between classes to form different semantic levels of classification models. Specifically, hierarchical clustering based on class relationships is used to form hierarchical clustering results, where the clustering levels are treated as semantic levels to form the classification models. Moreover, a new orthogonal module and a two-branch based CAM generation method are proposed to generate class regions that are orthogonal and complementary. We use the PASCAL VOC 2012 dataset to verify the proposed method. Experimental results show that our approach improves the CAM generation.