文献类型：期刊文献

英文题名：Rethinking Mobilevitv2 for Real-Time Semantic Segmentation

作者：Zhang, Zhengbin[1]; Xu, Zhenhao[1]; Gu, Xingsheng[1]; Xiong, Juan[2]

机构：[1] East China University of Science and Technology, School of Information Science and Engineering, Shanghai, China; [2] University of Shanghai for Science and Technology, School of Optical Electircal and Computer Engineering, Shanghai, China

年份：2023

起止页码：57

外文期刊名：2023 4th International Conference on Computers and Artificial Intelligence Technology, CAIT 2023

收录：EI(收录号：20241615928170)

语种：英文

外文关键词：Computer vision - Economic and social effects - Semantics

摘要：Semantic segmentation empowers various real-world applications. Nevertheless, the substantial computational cost, such as O(k2) time complexity associated with the number of tokens in multi-headed self-attention, poses challenges for deploying these models on edge devices with constrained hardware resources. This paper introduces a novel family of backbones designed for real-time semantic segmentation, referred to as Linear and Re-parameter Vision Transformer (LARFormer). Particularly, we introduce a Re-parameter Mobile Block (RMB), which employs three branches during training and a single branch during inference. Furthermore, we introduce Linear Separable Self-Attention(LSSA), which reduces the computational complexity from O(k2) to O(k). Extensive experiments on the ADE20K dataset and Pascal VOC 2012 dataset demonstrate the effectiveness of the proposed LARFormer by achieving a promising trade-off between segmentation accuracy and inference speed. ? 2023 IEEE.