Exploiting Features with Split-and-Share Module

TL;DR

This paper introduces the Split-and-Share Module (SSM), a novel classifier component that enhances CNN performance by encouraging shared features among sub-classifiers, validated through extensive experiments on ImageNet-1K.

Contribution

The paper proposes SSM, a new classifier module that can be integrated into existing CNNs to improve feature sharing and classification accuracy.

Findings

SSM improves baseline CNN architectures on ImageNet-1K.

SSM leads to more shared features among sub-classifiers.

Grad-CAM analysis shows better feature utilization with SSM.

Abstract

Deep convolutional neural networks (CNNs) have shown state-of-the-art performances in various computer vision tasks. Advances on CNN architectures have focused mainly on designing convolutional blocks of the feature extractors, but less on the classifiers that exploit extracted features. In this work, we propose Split-and-Share Module (SSM),a classifier that splits a given feature into parts, which are partially shared by multiple sub-classifiers. Our intuition is that the more the features are shared, the more common they will become, and SSM can encourage such structural characteristics in the split features. SSM can be easily integrated into any architecture without bells and whistles. We have extensively validated the efficacy of SSM on ImageNet-1K classification task, andSSM has shown consistent and significant improvements over baseline architectures. In addition, we analyze the effect of SSM using the Grad-CAM visualization.