基于HOG特征和SVM的手势识别

Hand Gesture Recognition Based on HOG Characters and SVM

手势识别具有比较丰富的应用领域,而领域的变化也会对识别的结果产生较大的影响。在诸多手势识别方法中,基于计算机视觉的识别技术对环境最敏感,例如光线亮暗、复杂背景、手势旋转等。为克服环境带来的影响,借鉴了近年来在目标检测研究中应用较多的梯度方向直方图技术,将其用于手势识别中。这种方法使得基于视觉的手势识别对环境不再敏感,得到了较好的识别效果。

Hand gesture recognition has abundant application fields, and the change of fields will result great influence to the result of recognition. In all gesture recognition methods, the one which based on computer vision is most sensitive to environment, such as too bright or too dark light, complex background, hand gesture rotation, etc. In order to over- come the impact of environment, histograms of oriented gradient which is being used in target detection research widely is used in hand gesture recognition. This method makes hand gesture recognition based on computer vision be sensitive to environment no longer, and a good recognition effect has been achieved.