说明：  本文提出一种用于手势识别的新方法，它将图像的方向直方图矢量(0Hv)与神经网络相结合。其特点在于选用 图像的方向直方图矢量作为手势的特征矢量，该特征矢量对于光线和手的平移变化具有较强的鲁棒性，这正是手势识别所要 解决的关键问题。在训练阶段，首先需要建立手势样本的特征矢量库；在识别阶段，本文选用三层BP网络作为分类器，获得 了90％以上的识别率。本文还对手势进行一定角度的旋转后的识别进行了讨论，识别结果达到预期要求。
(In this paper，a new method of gesture recognition is proposed，which combines orientation histogram vector(OHV) and neural net(NN)．Its high light is that OHV is selected as the feature vector of gestures，which is robust to illumination and translation and these two problems are very important in gesture recognition．In the training stage，a database of feature vectors is set up；In the recognition stage，a three layer BP neural net is selected tO classify gestures and the recognition rate is above 90％．The recognition of gestures after turning some angle is also discussed in this paper，the result is satisfying．)