Image-deblurring-

对图像灰度二值化，提取图像中某个色彩的部分并分割(The image is binarized and the part of a color in the image is extracted and segmented)

利用卷积神经网络进行特征提取，并将卷积特征可视化(Feature extraction using convolutional neural networks and visualization of convolution features)

最新图像轮廓透明度计算算法。 发表在今年的计算机视觉会议(latest image contours transparency calculation algorithm. Published in this year's Computer Vision Conference)

时频分布处理（重排、图像的信息处理以及Winger-Hough变换）(Time-Frequency Distribution Processing (rearrangements, the image information processing and the Hough transform Winger))

用光流法进行运动物体的检测和跟踪，并返回运动方向(Light flow method of moving object detection and tracking, and return to the direction of movement)

通论算法，给人脸做重构处理，进一步完成人脸的样本稀疏化，还能更好的重构人脸。(The General Theory of algorithms, giving face to do reconstruction processing to complete further thinning of face samples, but also to better reconstruct the face.)

说明：  对图像进行中值金字塔分解后，分别求取各层的概率密度函数(Image after median pyramid decomposition, respectively, to strike all floors of the probability density function)

图像处理中常用到的二十多个基本算法的源代码，如图像变换、灰度变换、图像滤波、锐化、边缘检测、阈值分割、边界提取、边界跟踪等(Commonly used in image processing to more than twenty basic algorithm source code, such as image conversion, gray-scale transformation, image filtering, sharpening, edge detection, threshold segmentation, edge detection, boundary tracking)

说明：  sprites communication to windows in unix sys

SVM 用于基于块划分特征提取的图像分类源码(SVM for feature extraction block-based image classification into source)