关键词: 梯度方向直方图, 支持向量机, 手势检测

Abstract:

In recent years,hand gesture detection becomes a research hotspot because of its broad application fields.In this paper,histograms of oriented gradient combined with support vector machine are used in hand gesture detection,which reduces the affect of illumination and rotation of gesture.HOG extraction and parameter settings are analyzed in detail,and a better classifier is trained based on HOG and linear SVM.Experiments show that the combination of HOG extraction and SWM can dramatically reduce the limitations on the application of hand gesture detection.

Key words: HOG;SVM;hand gesture detection