Compressive-Classifier

读取一个文件夹中的任意数量图片，并对读入图片进行尺寸调整、灰度化处理、直方图均衡化处理，并显示图片(Read number of images from a folder, and adjust the size,do gray processing, histogram equalization processing with the pictures , and display pictures.)

用差分盒维数处理图像，通过对图像分块，然后每一块来计算该区域的分维数，将其作为该区域所有像素上每一像素的分维数。然后，增强分维数分布图效果，然后取阈值，最后，给人造目标加框。(Differential box counting process the image, the image block, and then every one to calculate the fractal dimension of the region, as all the pixels in the region of each pixel on the fractal dimension. Then enhanced fractal dimension distribution effect, and then take the threshold, and finally, to the man-made objects framed.)

图像配准的很好的程序，里面用到粒子群算法(image registration is a good procedure, which used PSO)

matlab回归法做图像变化检测。有图片有结果，有论文说明，可参照使用(Matlab regression method for image change detection. There are pictures of the results, there are papers, can refer to use)

用于显著性检测的对比算法SP的MATLAB版本，比较新的是2013年的，亲测可用(MatLAB version of SP, a contrast algorithm for saliency detection, is relatively new in 2013, and pro-test is available.)

C# 实现验证码识别

/** * jpg图像文件缩放类 * * 本类实现一个对 JPG/JPEG 图像文件进行缩放处理的方法 即给定一个 JPG 文件，可以生成一个该 JPG 文件的缩影图像文件 (JPG 格式 ) *，(/ * * * JPG * * the zoom image file class to implement a class of JPG/JPEG image file to zoom method is given a JPG file, you can generate a miniature image file of the JPG file (JPG format) *,)

contourlet对LENA512灰度图的LP,DEF分解与系数重构代码,经过调试.(Contourlet LENA512 grayscale of the LP, DEF coefficient decomposition and reconstruction of the code after debugging.)

针孔模型中单应矩阵的求解 mian函数，测试单应矩阵计算的算法，程序流程如下 1、 读入两个图像。 2、读入4组对应点。 3、调用CalcH函数计算H矩阵。 4、在第一幅图中用鼠标选择一点。 5、调用WarpH函数计算在第二幅图中的对应的。 6、显示结果.(Pinhole model for solving the homography mian functions, test homography computation algorithm, program flow is as follows 1, read the two images. 2, 4 reads the corresponding points. 3, call CalcH function to calculate the H matrix. 4, in the first picture using the mouse to select it. 5, call WarpH function calculation in the second picture in the corresponding. 6 displays the results. )

噪声估计法，先定位噪声点，再实行去噪处理(Noise estimation method, first locate the noise points, then implement denoising.)