filter

用粒子滤波实现视频中人头跟踪算法，有训练好的人头分类器。()

传统的图像分割方法可以分为基于阈值、边缘、区域的方法等，而传统的分割方法在复杂图像分割中效果不好。为了提高图像分割的性能，我们提出了一种基于局部自适应窗口的Otsu方法。首先，分析了传统阈值法在复杂图像分割中不能很好地形成的原因，比较了全局阈值和局部阈值对图像分割的影响。其次，提出了根据局部信息自适应地改变局部窗口大小的方法，分析了该方法的特点。最后，证明了所提出的新方法优于其他方法。实验结果表明，与其他传统方法相比，该方法能保持更多的细节，获得更满意的结果。(The traditional image segmentation methods can be divided into thresholding, edge and region methods. The traditional segmentation methods do not perform well in the segmentation of complicated images. In order to improve the performance of image segmentation, we propose an adaptive windowed range-constrained Otsu method based on local information.First, the reason why traditional thresholding methods do not perform well in the segmentation of complicated images is analyzed. Therein, the influences of global and local thresholdings on the image segmentation are compared. Secondly, a method of adaptively changing the local window size based on local information is proposed, and the characteristics of the proposed method are analyzed. Finally, the superiority of the proposed method over other methods, It is validated by the experiments that the proposed method can keep more details and acquire much more satisfying results as compared with the other conventional methods.)

Retinex算法本质上就是一种基于光照补偿的图像增强算法, 能够对高动态范围图像进行增强。(Retinex algorithm is essentially an image enhancement algorithm based on illumination compensation, which can enhance the image of high dynamic range.)

说明：  基于HLS的图像处理，opencv，可供软件人员使用(Image processing based on HLS)

傅立叶变换，等距离采样，等角度采样，轮廓提取，形态处理(Fourier transform, equidistant sampling, in terms of sampling, contour extraction, morphological processing)

最大熵多阈值实现多红外图像的实时处理，多阈值实现(The maximum entropy multi-threshold to achieve real-time processing of multiple infrared image.)

VisualC++数字图像处理技术详解第2版光盘-第七章(VisualC++ digital image processing technology Detailed Version 2 CD- Chapter VII)

说明：  基于互信息的配准程序，matlab版的，内部包括测试图像为CT和ＭＲ。是老外编的，不错。( Matlab Program to find mutual information for two images stored as vectors. The images must contain 8-bit (0-255) integer pixels Requires two data vectors X,Y which must be the same length and Nx1 in size. Written by B. Corner--2003, University of Nebraska-Lincoln )

实现非均匀傅里叶变换代码，可用于石油地震勘探、航空航天、图像处理，电子电工方向。(Achieve non-uniform Fourier Transform code, can be used for seismic exploration, aerospace, image processing, electronics and electric direction.)

RPCA (Robust Principal Component Analysis)是目前用于矩阵填充、图像去噪的最有效的优化方法。目前最有效的算法是ALM（Augmented Lagrange Multiplier）。ALM分为Exact ALM和Inexact ALM。 该代码是Inexact ALM，收敛速度比Exact ALM快！(RPCA (Robust Principal Component Analysis) is used for matrix filling, image denoising. It is currently the most effective optimization method. Currently the most effective method is ALM (Augmented Lagrange Multiplier). There re 2 kinds of ALM: Exact ALM and Inexact ALM. The code is Inexact ALM, faster convergence speed than the Exact ALM!)