基于肤色模型和背景差分的手指区域分割方法

Finger Region Segmentation Based on Skin Colour Model and Background Subtraction

为了提高基于计算机视觉的人机交互系统中触控手指区域分割的性能和精度,提出了一种基于肤色模型和背景差分的手指区域分割方法。首先,根据肤色的聚类特征,在YCb Cr色彩空间上建立肤色模型,对手指区域进行肤色分割;随后加入自适应更新的背景差分法摒除图像中光照度和背景颜色的干扰。实验结果证明:该算法在复杂背景下分割手指区域具有较高的准确性和实时性,能应用在嵌入式环境中。

In order to improve the performance and accuracy of touched finger region segmentation in humancomputer interaction based on computer vision systems,this paper presented a finger region segmentation method based on skin color model and background difference. Firstly,according to the clustering character of skin color,the human hand was segmented via building the skin color model on the YCb Cr color space. Then by adding adaptive background subtraction method,the interference caused by illumination and background color was excluded. Experimental results show that the algorithm has higher accuracy and real-time ability under complex background. The proposed method can be applied in embedded environment.