Segmentation

基于Canny算子检测图像边缘，其中含有非极大值抑制功能。(Based on Canny operator edge detection image, which contains a non-maxima suppression.)

说明：  应用DCT字典和KSVD学习字典，对图像进行简单的修复，可行性较高，速度不是很快。(Simple image restoration is feasible and not very fast.)

C#连接摄像头，可以干好多事情了，自己扩展，这个只是个demo

将RGB格式图像转换到YUV空间，输入RGB图像返回YUV格式数据(Converting RGB format images to YUV space)

转换图象为二值图象后，进行识别与标记的matlab源码。可以计算目标物个数，形心等。(image conversion value for the two images, identification and marking of Matlab source. Can be calculated from the number of goals, such as heart-shaped.)

这是DBSCAN超像素分割的代码，可以发表个很好的论文，(This is a DBSCAN hyperpixel segmentation code that can publish a very good paper,)

使用snake方法进行图像处理的原代码，简单准确(use snake image processing methods for the original code simpler and more accurate)

正余弦+NLM滤波器和正余弦与其他滤波器的视觉效果与数值信息比较(Sine and Cosine+NLM cosine filter and compare numeric information with visual effects and other filters)

可用于视频目标检测中背景生成的混合高斯模型(Can be used for video object detection in the context of generating a mixture of Gaussian model)

可以实现平面图像到球面坐标的映射变换，效果较好，附原始图片。(Plane image mapping transformation to spherical coordinates better with the original picture.)