Global Self-Attention Networks for Image Recognition

Abstract

Recently, a series of works in computer vision have shown promising results on various image and video understanding tasks using self-attention. However, due to the quadratic computational and memory complexities of self-attention, these works either apply attention only to low-resolution feature maps in later stages of a deep network or restrict the receptive field of attention in each layer to a small local region. To overcome these limitations, this work introduces a new global self-attention module, referred to as the GSA module, which is efficient enough to serve as the backbone component of a deep network. This module consists of two parallel layers: a content attention layer that attends to pixels based only on their content and a positional attention layer that attends to pixels based on their spatial locations. The output of this module is the sum of the outputs of the two layers. Based on the proposed GSA module, we introduce new standalone global attention-based deep networks that use GSA modules instead of convolutions to model pixel interactions. Due to the global extent of the proposed GSA module, a GSA network has the ability to model long-range pixel interactions throughout the network. Our experimental results show that GSA networks outperform the corresponding convolution-based networks significantly on the CIFAR-100 and ImageNet datasets while using less parameters and computations. The proposed GSA networks also outperform various existing attention-based networks on the ImageNet dataset.

References in corpus (7)

• Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
• Large Scale GAN Training for High Fidelity Natural Image Synthesis
• Self-Attention Generative Adversarial Networks
• Axial Attention in Multidimensional Transformers
• VideoBERT: A Joint Model for Video and Language Representation Learning
• Compact Generalized Non-local Network
• Axial-DeepLab: Stand-Alone Axial-Attention for Panoptic Segmentation

Cited by in corpus (5)

• Revisiting ResNets: Improved Training and Scaling Strategies
• LeViT: a Vision Transformer in ConvNet's Clothing for Faster Inference
• Interpretable CNN-Multilevel Attention Transformer for Rapid Recognition of Pneumonia from Chest X-Ray Images
• Generative Adversarial Transformers
• A Comparison for Anti-noise Robustness of Deep Learning Classification Methods on a Tiny Object Image Dataset: from Convolutional Neural Network to Visual Transformer and Performer