摘要
采用隐马尔可夫模型(Hidden Markov Model)训练并识别手势样本,针对HMM的经典训练算法Baum-Welch算法的缺点,采用纠错算法对其修正,提高了识别率。将识别结果应用于“基于Internet远程机器人控制”项目,达到了对机器人控制的目的,优化了人机交互的接口。
The Hidden Markov Model (HMM) was used and the rectify method was also introduced to correct the training of HMM, and a better result was gotten. Finally, the results of dynamic hand gesture recognition were introduced to the “Internet based remote Robot control”project, which realized the objective of controlling robot and optimized the Human Computer Interactive interface,
出处
《计算机工程与设计》
CSCD
北大核心
2005年第11期2934-2936,共3页
Computer Engineering and Design
基金
上海市科委重点攻关基金项目(015115042)
关键词
隐马尔可夫模型
模式识别
光流跟踪
手势分割
阈值
HMM
pattern recognition
optical flow tracking
hand gesture extraction
threshold