Parameter-Free Channel Attention for Image Classification and Super-Resolution

Yuxuan Shi,Lingxiao Yang,Wangpeng An,Xiantong Zhen,Liuqing Wang

2023-03-20

Abstract:The channel attention mechanism is a useful technique widely employed in deep convolutional neural networks to boost the performance for image processing tasks, eg, image classification and image super-resolution. It is usually designed as a parameterized sub-network and embedded into the convolutional layers of the network to learn more powerful feature representations. However, current channel attention induces more parameters and therefore leads to higher computational costs. To deal with this issue, in this work, we propose a Parameter-Free Channel Attention (PFCA) module to boost the performance of popular image classification and image super-resolution networks, but completely sweep out the parameter growth of channel attention. Experiments on CIFAR-100, ImageNet, and DIV2K validate that our PFCA module improves the performance of ResNet on image classification and improves the performance of MSRResNet on image super-resolution tasks, respectively, while bringing little growth of parameters and FLOPs.

Image and Video Processing,Computer Vision and Pattern Recognition

What problem does this paper attempt to address?

The paper mainly aims to address the following two issues:

1. **Reducing the parameter count and computational cost of channel attention mechanisms**: Existing channel attention mechanisms can effectively improve the performance of image classification and super-resolution tasks, but they introduce additional parameters, leading to increased computational costs. Therefore, the paper proposes a Parameter-Free Channel Attention (PFCA) module, which aims to maintain or enhance the performance of existing networks while avoiding the addition of extra parameters.

2. **Application to image classification and super-resolution tasks**: The PFCA module is designed as a plug-and-play component that can be embedded into existing image classification and super-resolution networks to enhance their feature representation capabilities. Specifically, the PFCA module is integrated into the ResNet series networks (for image classification tasks) and MSRResNet (for image super-resolution tasks) and has been validated on multiple benchmark datasets.

Through the above work, the paper demonstrates the effectiveness of the PFCA module, showing that it not only improves the performance of ResNet and MSRResNet networks but also adds almost no extra computational cost and parameter count.