Channel Locality Block: A Variant of Squeeze-and-Excitation

TL;DR

This paper introduces the Channel Locality (C-Local) block, a variant of the Squeeze-and-Excitation block that uses convolutional layers to capture local channel dependencies, improving CNN performance.

Contribution

The paper proposes the C-Local block, replacing full connection layers with convolutional layers to model local channel correlations in attention mechanisms.

Findings

C-Local block achieves higher accuracy than SE block on CIFAR-10.

C-Local effectively captures local channel dependencies.

Improved CNN performance with the new attention module.

Abstract

Attention mechanism is a hot spot in deep learning field. Using channel attention model is an effective method for improving the performance of the convolutional neural network. Squeeze-and-Excitation block takes advantage of the channel dependence, selectively emphasizing the important channels and compressing the relatively useless channel. In this paper, we proposed a variant of SE block based on channel locality. Instead of using full connection layers to explore the global channel dependence, we adopt convolutional layers to learn the correlation between the nearby channels. We term this new algorithm Channel Locality(C-Local) block. We evaluate SE block and C-Local block by applying them to different CNNs architectures on cifar-10 dataset. We observed that our C-Local block got higher accuracy than SE block did.