# Cross Attention Network for Semantic Segmentation

## TL;DR

This paper introduces a Cross Attention Network for semantic segmentation that effectively combines low-level spatial details with high-level contextual features using a novel FCA module, achieving superior accuracy and speed.

## Contribution

The paper proposes a new FCA module that separately fuses spatial and channel attention from two network branches, enhancing segmentation performance.

## Key findings

- Outperforms existing real-time methods on Cityscapes and CamVid datasets.
- Achieves state-of-the-art accuracy with deep backbones.
- Maintains high speed with lightweight backbones.

## Abstract

In this paper, we address the semantic segmentation task with a deep network that combines contextual features and spatial information. The proposed Cross Attention Network is composed of two branches and a Feature Cross Attention (FCA) module. Specifically, a shallow branch is used to preserve low-level spatial information and a deep branch is employed to extract high-level contextual features. Then the FCA module is introduced to combine these two branches. Different from most existing attention mechanisms, the FCA module obtains spatial attention map and channel attention map from two branches separately, and then fuses them. The contextual features are used to provide global contextual guidance in fused feature maps, and spatial features are used to refine localizations. The proposed network outperforms other real-time methods with improved speed on the Cityscapes and CamVid datasets with lightweight backbones, and achieves state-of-the-art performance with a deep backbone.