现代电子技术2025，Vol.48Issue(23)：151-154,4.DOI:10.16652/j.issn.1004-373x.2025.23.022
考虑光流匹配的人机交互动态手势识别方法
Dynamic gesture recognition method for human-computer interaction considering optical flow matching
摘要
Abstract
To enhance adaptability to real interactive scenarios and meet the recognition challenges caused by changes in gesture speed and partial occlusion during human-computer interaction(HCI),a dynamic gesture recognition method considering optical flow matching is proposed.In the method,an optical flow algorithm improved by Gaussian pyramid multi-resolution structure is used to track gesture motion trajectories,generate gesture trajectory optical flow feature sequences,and align them with the gesture optical flow model sequences on the time axis by the dynamic time warping(DTW)algorithm.The dynamic gesture recognition is achieved by calculating distortion and comparing it with a set threshold value.The experimental results show that the proposed method can achieve accurate recognition of different dynamic gestures,with F1 score and FPS of 0.91 and 40.8 f/s,respectively,and its overall performances are better than those of the comparative methods.To sum up,the proposed method provides a high-quality interactive scheme for digital art creation and experience.关键词
光流匹配/人机交互/动态手势识别/肤色模型/Gaussian金字塔/动态时间规整算法Key words
optical flow matching/HCI/dynamic gesture recognition/skin color model/Gaussian pyramid/DTW algorithm分类
信息技术与安全科学引用本文复制引用
张建淳,张博尊..考虑光流匹配的人机交互动态手势识别方法[J].现代电子技术,2025,48(23):151-154,4.基金项目
2022年度黑龙江省艺术科学规划项目(2022A017) （2022A017）
