基于唇形的智能轮椅人机交互

Human-machine Interaction Based on Shape of Lip for Intelligent Wheelchair

提出了一种新颖的基于唇形的智能轮椅人机交互方法。该方法通过定义不同的唇形来控制智能轮椅的运动,使用者在噪杂环境中也能进行和谐、方便地人机交互。首先使用adaboost算法在视频帧中实时、准确地检测唇部,然后通过离散余弦变换(Discrete Cosine Transform DCT)提取唇部特征。针对经DCT变换后特征矢量维数较高以及支持向量机(Support Vector Machine SVM)在解决小样本、非线性及高维模式识别中表现出的优势,最后使用SVM进行唇形的识别,并将识别结果转换成控制指令来控制轮椅的运动。实验证明,该方法能有效地识别不同唇形并可用于实时控制智能轮椅的运动。

The movement of intelligent wheelchair can be controlled by changing different shape of lip. Users can communicate with the intelligent wheelchair harmoniously and expediently even in the noise environment. The adaboost algorithm is used to detect the lip area in the real-time video frame and the features of lip are extracted by Discrete Cosine Transform （DCT）. Owing to the dimension of feature after DCT is very high and the Support Vector Machine （SVM） shows great advantage in small sample, non-linear, pattern recognition in high dimension space. So, SVM is used to recognize the shape of lip. The recognition results are converted to the commands of intelligent wheelchair for the movement controlling. The experimental results show that this method can recognize the shapes of lip in different lighting environment and control the movement of intelligent wheelchair effectively.