ELA: Efficient Local Attention for Deep Convolutional Neural Networks

arXiv:2403.01123v1 Announce Type: new
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 …

abstract arxiv attention computer computer vision convolutional neural networks cost cs.cv information local attention networks neural networks performance recognition spatial struggle type vision