STEAM: Squeeze and Transform Enhanced Attention Module

TL;DR

This paper introduces STEAM, a novel graph-based attention module that models both channel and spatial attention efficiently, improving CNN performance with minimal additional computational cost.

Contribution

The paper proposes the first graph-based approach to jointly model channel and spatial attention with constant parameters, including a new pooling method for spatial context.

Findings

Achieves 2% higher accuracy on image classification with minimal GFLOPs increase.

Outperforms ECA and GCT modules in accuracy while reducing GFLOPs threefold.

Effective for object detection and segmentation tasks.

Abstract

Channel and spatial attention mechanisms introduced by earlier works enhance the representation abilities of deep convolutional neural networks (CNNs) but often lead to increased parameter and computation costs. While recent approaches focus solely on efficient feature context modeling for channel attention, we aim to model both channel and spatial attention comprehensively with minimal parameters and reduced computation. Leveraging the principles of relational modeling in graphs, we introduce a constant-parameter module, STEAM: Squeeze and Transform Enhanced Attention Module, which integrates channel and spatial attention to enhance the representation power of CNNs. To our knowledge, we are the first to propose a graph-based approach for modeling both channel and spatial attention, utilizing concepts from multi-head graph transformers. Additionally, we introduce Output Guided Pooling (OGP), which efficiently captures spatial context to further enhance spatial attention. We extensively evaluate STEAM for large-scale image classification, object detection and instance segmentation on standard benchmark datasets. STEAM achieves a 2% increase in accuracy over the standard ResNet-50 model with only a meager increase in GFLOPs. Furthermore, STEAM outperforms leading modules ECA and GCT in terms of accuracy while achieving a three-fold reduction in GFLOPs.