Full body movements recognition - unsupervised learning approach with heuristic R-GDL method
详细信息 查看全文
- 作者：Tomasz Hachaja ; tomekhachaj@o2.pl" class="auth_mail" title="E-mail the corresponding authorAuthor Vitae ; Marek R. Ogielab ; mogiela@agh.edu.pl" class="auth_mail" title="E-mail the corresponding authorAuthor Vitae
- 关键词：Gesture recognition ; Full body movements recognition ; Unsupervised learning ; Syntactic classification ; Gesture Description Language ; Reverse-Gesture Description Language
- 刊名：Digital Signal Processing
- 出版年：2015
- 出版时间：November 2015
- 年：2015
- 卷：46
- 期：Complete
- 页码：239-252
- 全文大小：1554 K
文摘
- •
We have used Gesture Description Language (GDL) for user actions recognition.
- •
A method for generation of knowledge base for GDL classifier has been proposed.
- •
The method analyzes unsegmented data recordings of gestures.
- •
The evaluation has been made on 770 samples of gym exercises.
- •
We obtained recognition rate at the level of 100% to 91%.