In the study of gesture recognition, it is often affected by changes in illumination, complex background, and skin color interference. In response to these problems, this paper designed a Kinect-based gesture recognition method. Based on the Kinect somatosensory device, this paper collected the bone information and binary image of the human body, and combines the position of the bone points to obtain the hand image. Then, the feature of the hand image was extracted using the HOG algorithm, and the gesture image was classified using the SVM algorithm. In order to implement this method, we collected 10 digital gestures through Kinect and evaluated the method through experiments. The experimental results show that based on the Kinect somatosensory device, the HOG algorithm and the SVM algorithm can effectively and efficiently recognize gestures, which significantly improves the accuracy of gesture recognition, and the average recognition rate is as high as 98.3%. This is of great significance for the research and promotion of gesture recognition technology.
[1]
Thomas B. Moeslund,et al.
Recognition of Deictic Gestures for Wearable Computing
,
2005,
Gesture Workshop.
[2]
Jovan Popović,et al.
Real-time hand-tracking with a color glove
,
2009,
SIGGRAPH 2009.
[3]
Yuji Iwahori,et al.
A HOG-SVM Based Fall Detection IoT System for Elderly Persons Using Deep Sensor
,
2018,
IIKI.
[5]
Antti Oulasvirta,et al.
Interactive Markerless Articulated Hand Motion Tracking Using RGB and Depth Data
,
2013,
2013 IEEE International Conference on Computer Vision.
[6]
Rytis Maskeliunas,et al.
Recognition of basketball referee signals from videos using Histogram of Oriented Gradients (HOG) and Support Vector Machine (SVM)
,
2018,
ANT/SEIT.