L1 Total Variation

实现对鱼眼图像的校正 我自己执行过 可以使用(it can realise the rectification of the fish eye image )

基于大津法和迭代法进行图像分割，分割结果较好(image segmentation based on otsu and interation)

说明：  语义分割，训练样本，模型训练，成果展示，深度卷积神经网络(Semantic segmentation, training samples, model training, achievement display, deep convolution neural network)

本文提出了一种新的快速自适应加权中值滤波算法以提高中值滤波算法的性能。(A new fast adaptive weighted median filtering algorithm is proposed to improve the performance of the median filtering algorithm.)

说明：  实现了图像的打开功能，并能实现图像打开时的扫描，渐进，马赛克等功能(To realize the image of the open function, and can achieve image scans open, progressive, mosaic and other functions)

凸多边形检测和显示--先把图像中的显著区域提取出来，然后用最小凸多边形囊括显著区域，并把凸多边形显示出来(Convex polygon detection and show- first images extracted significant area, and then use the smallest convex polygon covering a significant area, and the convex polygon displayed)

利用SFS算法实现3D图像重建，附带有图片实现(3D Image Reconstruction Using SFS Algorithms)

图像检索——提取颜色特征：HSV空间颜色直方图(将RGB空间转化为HSV空间并进行非等间隔量化，将三个颜色分量表示成一维矢量，再计算其直方图作为颜色特征。(Image Retrieval- extract color features: HSV space color histogram (RGB space into the HSV space and the non-interval to quantify the three color components expressed as a one-dimensional vector, then calculate the color histogram as a feature.)

盲信号分离应用的图像盲分离的资料，给出了事例以及部分源码(Blind Signal Separation Application Blind Image Separation of information, and gives examples of some of the FOSS)

说明：  高斯尺度差角点检测算法 将轮廓曲线在不同尺度下的演化差异的变化量作为角点响应函数并提出相应的 DoG 检测算子(Gauss scale difference corner detection algorithm)