ELA: Efficient Local Attention for Deep Convolutional Neural Networks

TL;DR

The paper introduces ELA, a lightweight local attention module that enhances deep CNNs' performance across various vision tasks by efficiently encoding positional information without increasing complexity.

Contribution

It proposes a novel ELA method that overcomes limitations of existing attention mechanisms using 1D convolution and Group Normalization, with multiple adaptable versions for different tasks.

Findings

ELA outperforms state-of-the-art methods on ImageNet, MSCOCO, and Pascal VOC datasets.

ELA improves accuracy in image classification, object detection, and semantic segmentation.

The method is compatible with popular CNN architectures like ResNet, MobileNet, and DeepLab.

Abstract

The attention mechanism has gained significant recognition in the field of computer vision due to its ability to effectively enhance the performance of deep neural networks. However, existing methods often struggle to effectively utilize spatial information or, if they do, they come at the cost of reducing channel dimensions or increasing the complexity of neural networks. In order to address these limitations, this paper introduces an Efficient Local Attention (ELA) method that achieves substantial performance improvements with a simple structure. By analyzing the limitations of the Coordinate Attention method, we identify the lack of generalization ability in Batch Normalization, the adverse effects of dimension reduction on channel attention, and the complexity of attention generation process. To overcome these challenges, we propose the incorporation of 1D convolution and Group Normalization feature enhancement techniques. This approach enables accurate localization of regions of interest by efficiently encoding two 1D positional feature maps without the need for dimension reduction, while allowing for a lightweight implementation. We carefully design three hyperparameters in ELA, resulting in four different versions: ELA-T, ELA-B, ELA-S, and ELA-L, to cater to the specific requirements of different visual tasks such as image classification, object detection and sementic segmentation. ELA can be seamlessly integrated into deep CNN networks such as ResNet, MobileNet, and DeepLab. Extensive evaluations on the ImageNet, MSCOCO, and Pascal VOC datasets demonstrate the superiority of the proposed ELA module over current state-of-the-art methods in all three aforementioned visual tasks.