Cross-CBAM: A Lightweight network for Scene Segmentation

TL;DR

The paper introduces Cross-CBAM, a lightweight real-time scene segmentation network that combines novel attention modules and multiscale pooling to achieve high accuracy and speed on edge devices.

Contribution

It proposes the SE-ASPP and CCBAM modules, enabling efficient multiscale feature extraction and feature fusion with cross-attention for improved real-time segmentation.

Findings

Achieves 73.4% mIoU at 240.9FPS on Cityscapes

Attains 77.2% mIoU at 88.6FPS on Cityscapes with GTX 1080Ti

Demonstrates a favorable accuracy-speed trade-off on benchmark datasets

Abstract

Scene parsing is a great challenge for real-time semantic segmentation. Although traditional semantic segmentation networks have made remarkable leap-forwards in semantic accuracy, the performance of inference speed is unsatisfactory. Meanwhile, this progress is achieved with fairly large networks and powerful computational resources. However, it is difficult to run extremely large models on edge computing devices with limited computing power, which poses a huge challenge to the real-time semantic segmentation tasks. In this paper, we present the Cross-CBAM network, a novel lightweight network for real-time semantic segmentation. Specifically, a Squeeze-and-Excitation Atrous Spatial Pyramid Pooling Module(SE-ASPP) is proposed to get variable field-of-view and multiscale information. And we propose a Cross Convolutional Block Attention Module(CCBAM), in which a cross-multiply operation is employed in the CCBAM module to make high-level semantic information guide low-level detail information. Different from previous work, these works use attention to focus on the desired information in the backbone. CCBAM uses cross-attention for feature fusion in the FPN structure. Extensive experiments on the Cityscapes dataset and Camvid dataset demonstrate the effectiveness of the proposed Cross-CBAM model by achieving a promising trade-off between segmentation accuracy and inference speed. On the Cityscapes test set, we achieve 73.4% mIoU with a speed of 240.9FPS and 77.2% mIoU with a speed of 88.6FPS on NVIDIA GTX 1080Ti.