arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2412.08670cs.CVcs.LGeess.IV
收藏

一种面向轻量级语义分割网络的特征细化模块

A feature refinement module for light-weight semantic segmentation network

School of Electrical and Information Engineering, Zhengzhou University(郑州大学电气与信息工程学院)

Zhiyan Wang, Xin Guo, Song Wang, Peixiao Zheng, Lin Qi

AI总结:

针对轻量级语义分割网络因参数量少、表征能力不足导致精度下降的问题,提出含特征细化模块的新方法,通过多阶段特征提取与transformer捕获非局部上下文,在Cityscapes等数据集上实现精度与计算成本的良好权衡。

摘要:

低计算复杂度与高分割精度对于实际场景下的语义分割任务均至关重要。然而,为加快模型推理速度,现有多数方法往往设计参数量极为有限的轻量级网络,这会因网络表征能力下降导致精度大幅降低。为解决该问题,本文提出一种新型语义分割方法,以提升轻量级网络的语义信息获取能力。具体而言,本文提出特征细化模块(FRM),该模块从骨干网络生成的多阶段特征图中提取语义信息,并利用transformer模块捕获非局部上下文信息。在Cityscapes和Bdd100K数据集上的实验结果表明,所提方法在精度与计算成本之间实现了良好的权衡,尤其在Cityscapes测试集上达到80.4%的mIoU,仅需214.82 GFLOPs的计算量。

英文摘要:

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.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑