A feature refinement module for light-weight semantic segmentation network

TL;DR

This paper introduces a feature refinement module that enhances semantic information extraction in light-weight segmentation networks, balancing accuracy and computational efficiency.

Contribution

It proposes a novel feature refinement module utilizing transformer blocks to improve semantic capacity in light-weight networks.

Findings

Achieves 80.4% mIoU on Cityscapes with low computational cost

Outperforms existing light-weight models in accuracy-cost trade-off

Demonstrates effectiveness on Bdd100K dataset

Abstract

Low computational complexity and high segmentation accuracy are both essential to the real-world semantic segmentation tasks. However, to speed up the model inference, most existing approaches tend to design light-weight networks with a very limited number of parameters, leading to a considerable degradation in accuracy due to the decrease of the representation ability of the networks. To solve the problem, this paper proposes a novel semantic segmentation method to improve the capacity of obtaining semantic information for the light-weight network. Specifically, a feature refinement module (FRM) is proposed to extract semantics from multi-stage feature maps generated by the backbone and capture non-local contextual information by utilizing a transformer block. On Cityscapes and Bdd100K datasets, the experimental results demonstrate that the proposed method achieves a promising trade-off between accuracy and computational cost, especially for Cityscapes test set where 80.4% mIoU is achieved and only 214.82 GFLOPs are required.