程序

File from feature-based image metamorphosis

改进的聚类光流法，聚类方法有kmeans和DBSCAN，光流法金字塔LK可以用，HS光流法占用内存太多。(Improved clustering optical flow method, clustering method kmeans and DBSCAN, pyramid LK optical flow method can be used, HS optical flow occupy too much memory.)

说明：  图像增强的目标是改进图片的质量，例如增加对比度，去掉模糊和噪声，修正几何畸变等；图像复原是在假定已知模糊或噪声的模型时，试图估计原图像的一种技术。 图像增强按所用方法可分成频率域法和空间域法。前者把图像看成一种二维信号，对其进行基于二维傅里叶变换的信号增强。采用低通滤波（即只让低频信号通过）法，可去掉图中的噪声；采用高通滤波法，则可增强边缘等高频信号，使模糊的图片变得清晰。具有代表性的空间域算法有局部求平均值法和中值滤波（取局部邻域中的中间像素值）法等，它们可用于去除或减弱噪声。 (A sharp and clear image is not directly produced from the optics, however, digital signal processing applied to the sampled image produces a sharp and clear final image that is also insensitive to misfocus related aberrations. This paper gives an overview of Wavefront Coding and example images related to the two applications of machine vision/label reading and biometric imaging. Design techniques of Wavefront Coding are unique from that of traditional imaging system design since both the optics and digital processing characteristics of the system are jointly optimized for optimum system performance.)

图像融合程序，及论文 图像融合程序，及论文 图像融合程序，及论文(Image fusion procedure, and the paper image fusion procedure, and the paper image fusion process, and papers)

去除高斯噪声的有效码源，去噪能力强，使简单。(The effective removal of Gaussian noise source code, denoising capability, so that simple.)

说明：  如题，matlab手写数字测试图像处理 样本（30+50个）加全代码加测试结果(Like title, matlab handwritten digital test image processing samples (30 + 50) plus full code plus test results)

图像去噪，包括小波去噪、均值滤波、中值滤波、自适应维纳滤波等(Image denoising)

some fast feature extraction algorithms are addressed for joint retrieval of images compressed in JPEG and JPEG2000 formats. In order to avoid full decoding, three fast algorithms that convert block-based discrete cosine transform (BDCT) into wavelet transform are developed, so that wavelet-based features can be extracted from JPEG images as in JPEG2000 images.

基于形态学的滤波算法，tophat算法的matlab源代码即图片(Filtering algorithm based on morphology, tophat algorithm matlab source code for the image)

这是用C++实现图像立体匹配的一个完整的论文代码，主要特性是提出了一种全新的跨规模代价聚合方法，适合对于图像的立体匹配有一定基础的人学习(Cross-Scale Cost Aggregation for Stereo Matching)