这是利用神经网络来实现手写字符识别，起准确率已经达到99.26 ，可以继续调整参数达到更深层次的效果。需要自己搭建opencv环境。后期工作可以利用cuda对其更深层次的加速-This is achieved using a neural network handwritten character recognition, since the exact rate has reached 99.26 percent, can continue to adjust the parameters t

利用cuda+opencv对照片进行高斯模糊-image blur using cuda and opencv

求深度信息的半全局匹配算法，使用cuda，opencv(For depth information of the semi global matching algorithm, using CUDA, opencv)

实现视频实时拼接，基于opencv和cuda编程，能实现实时行(Real-time video splicing, based on OpenCV and CUDA Programming, can realize real time line)