- 期刊
Dynamic Gesture Recognition Method Based on Conditional Random Field and Weighted Voting Strategy
摘要
Aiming at the diversity and spatio-temporal difference of dynamic gestures, the over-fitting and feature interference caused by the increase of feature dimensions, this paper proposes a dynamic gesture recognition method based on conditional random field and weighted voting method. Firstly, the two kinds of features of hand shape and trajectory are extracted. Then, the conditional random field model is used to model different feature subsets to obtain the difference base classifier. Finally, using the weighted voting method to fuse the base classifier to judge the gesture. The experimental results show that the method can accurately identify the hand shape and position with rich dynamic changes.
延伸閱讀
- 洪偉哲（2014）。Development and Assessmen of Instant Gesture Recognition Algorithm Based on Wearable Device〔碩士論文，國立臺北科技大學〕。華藝線上圖書館。https://doi.org/10.6841/NTUT.2014.00226
- 洪兆欣（2006）。A Trajectory-based Approach to Gesture Recognition〔碩士論文，國立中央大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0031-0207200917342420
- 陳彥吉（2016）。A Depth Image-based Fuzzy Hand Gesture Recognition Method〔碩士論文，朝陽科技大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0078-1108201714030875
- 林本煌（2011）。Gesture Recognition Using Improved Hierarchical Hidden Markov Model〔碩士論文，國立臺北科技大學〕。華藝線上圖書館。https://doi.org/10.6841/NTUT.2011.00714
- 呂緯昱（2010）。A Gesture Recognition Method based on the Finger Tips and Angles with Local Skin Color Detection〔碩士論文，國立臺北科技大學〕。華藝線上圖書館。https://doi.org/10.6841/NTUT.2010.00129