基于二级匹配策略的实时动态手语识别-计算机科学.PDFVIP

Vo1. 44 No. 7 第 44 卷第7 期 计算机科学 2017 年 7 月 COMPUTER SCIENCE July 2017 基于二级匹配策略的实时动态手语识别 梁文乐1 黄元元1 胡作进2 (南京航空航天大学计算机科学与技术学院 南京 210016)1 (南京特殊教育师范学院数学与信息科学学院 南京 210038)2 摘 要 动态手语可以利用其轨迹与关键手型加以描述。大量的统计实验数据表明，大多数的常用手语通过轨迹曲 线的匹配即可实现识别，因此，提出一种针对动态手语的分级匹配识别算法。首先利用体感设备获取手势轨迹，并根 据轨迹的点密度分布设计了一种关键帧检测算法以才是取手势的关键手型，结合轨迹的曲线特征，实现对动态手语的精 确描述。然后利用优化的动态时间规整(DTW) 算法完成对手语的一级匹配，即轨迹匹配。若此时可以得到识别结 果，那么识别过程可以结束，否则进入二级匹配，即针对关键手型再做匹配识别，从而得到最终的识别结采。实验证 明，所提算法不仅实时性好，识别的准确率也较高。 关键词 动态手语识别，手语轨迹，关键帧，动态时间规整 中图法分类号 TP391 文献标识码 A 001 10. 11896/j. issn. 1002-137叉 2017.07.054 Real-time Dynamic Sign Language Recognition Based on Hierarchical Matching Strategy LIANG Wen-le 1 HUANG Yuan-yuan1 HU Zuo-jin2 (College of Computer Science and Technology ,Nanjing University of Aeronautics and Astronautics ,Nanjing 210016 ,China) 1 (School of Mathematics and Information Science ,Nanjing Normal University of Special Education ,Nanjing 210038 ,China) 2 Abstract Dynamic sign language can be described by its trajectory and the key hand-action. However ,a large number of statistical data show that most of the commonly used sign languages can be recognized by its trajectory curve. There fore ,a hierarchical matching recognition strategy for dynamic sign language was proposed in this paper. First ,the ges ture trajectory can be obtained by the somatosensory equipment like Kinect. According to its point density ,an algorithm of key frame detection is designed and is used to extract the key gestures. Thus ,we can achieve a precise description of