Informative Class Activation Maps
This work addresses the need for more informative feature localization in image classification, particularly for weakly supervised tasks, though it appears incremental as it builds on existing class activation map methods.
The paper tackles the problem of evaluating the information content of image regions for a given label by bridging class activation maps with information theory, resulting in state-of-the-art performance on Tiny-ImageNet for weakly supervised object localization.
We study how to evaluate the quantitative information content of a region within an image for a particular label. To this end, we bridge class activation maps with information theory. We develop an informative class activation map (infoCAM). Given a classification task, infoCAM depict how to accumulate information of partial regions to that of the entire image toward a label. Thus, we can utilise infoCAM to locate the most informative features for a label. When applied to an image classification task, infoCAM performs better than the traditional classification map in the weakly supervised object localisation task. We achieve state-of-the-art results on Tiny-ImageNet.