采用深度图像信息和SLVW的手语识别

Improved sign language recognition research using depth image information and SLVW

为了实现手语视频中手语字母的准确识别,提出了一种基于DI_CamShift和SLVW的算法。该方法将Kinect作为手语视频采集设备,在获取彩色视频的同时得到其深度信息;计算深度图像中手语手势的主轴方向角和质心位置,通过调整搜索窗口对手势进行准确跟踪;使用基于深度积分图像的Ostu算法分割手势,并提取其SIFT特征;构建了SLVW词包作为手语特征,并用SVM进行识别。通过实验验证该算法,其单个手语字母最好识别率为99.87%,平均识别率96.21%。

In order to realize the accurate recognition of manual alphabets in the sign language video, this paper presents an im-proved algorithm based on DI_CamShift（Depth Image CamShift）and SLVW（Sign Language Visual Word）. It uses Kinect as the sign language video capture device to obtain both of the color video and depth image information of sign language gestures. The paper calculates spindle direction angle and mass center position of the depth images to adjust the search window and for gesture tracking. An Ostu algorithm based on depth integral image is used to gesture segmentation, and the SIFT features are ex-tracted. It builds the SLVW bag of words as the feature of sign language and uses SVM for recognition. The best recognition rate of single manual alphabet can reach 99.87%, and the average recognition rate is 96.21%.