计算机与数字工程2026，Vol.54Issue(2)：429-434,6.DOI:10.3969/j.issn.1672-9722.2026.02.022
面向工业人体轮廓遮挡场景的步态识别研究
Research on Gait Recognition in Industrial Environments with Human Silhouette Occlusions
李红江 1辛煌炜 2胡锦晖3
作者信息
- 1. 武汉蓝海科创技术有限公司 武汉 430035
- 2. 华中科技大学计算机科学与技术学院 武汉 430074
- 3. 武汉东湖学院机电工程学院 武汉 430212
- 折叠
摘要
Abstract
As a dynamic biometric technology,gait recognition has unique advantages that traditional static biometric technol-ogy does not possess,and has become one of the hot topics in current academic research.In industrial scenarios,gait recognition is suitable for identifying and monitoring the identity of individuals who exhibit abnormal behavior over medium to long distances,en-suring the safety of the recognition process without physical contact.However,people in industrial environments often carry a vari-ety of complex items,which can cause obstructions to the body,especially to the upper body,changing the contour characteristics of the human body and affecting the accuracy of gait recognition.An adaptive occlusion processing gait recognition network based on dual-segment segmentation is designed to address occlusion issues in industrial environments.The network includes a primary con-tour feature extraction module,a global feature extraction unit,a multi-scale micro action extraction module,and a feature fusion module based on adaptive occlusion perception.The experimental results show that the network has good recognition accuracy for large-scale occlusion cases.关键词
步态识别/卷积神经网络/特征提取/遮挡处理Key words
gait recognition/convolutional neural network/feature extraction/occlusion processing分类
信息技术与安全科学引用本文复制引用
李红江,辛煌炜,胡锦晖..面向工业人体轮廓遮挡场景的步态识别研究[J].计算机与数字工程,2026,54(2):429-434,6.
