content-based-image-retrieval

这是一个结合全局-局部图像对比度增强程序。首先使用全局对比度增强，然后使用局部对比度增强。程序中的参数是经过导师修改过的。文件中给出了原论文。对图像有一定的增强效果，仅供参考。(This is a combination of a global- local image contrast enhancement program. First, using a global contrast enhancement, and then use the local contrast enhancement. Program parameters are modified through mentor. Shows the original paper document. There is a certain image enhancement for reference purposes only.)

说明：  使用unet对图像进行分割的源码，里面有训练集，可以根据自己的需要更换训练数据。(Use the source code of the image segmentation using UNET, which has a training set, you can change the training data according to your own needs.)

YUV多种图像处理方式，包括yuv2rgb,rgb2yuv.等等。但需要仔细调试(YUV multiple image processing methods, including yuv2rgb, rgb2yuv. And so on. But needs to carefully debugging)

对采样得到的180幅断层图像进行重建，并针对其中的一幅图像进行处理，并且统计图像中颗粒的大小和数目！(Tomographic images of 180 samples obtained were rebuilt, and for which an image processing and image size and number of particles statistics!)

掌纹识别代码，内有详细代码以及train的图(Palmprint recognition code, there are detailed code and train diagram.cation code)

区域生长法的图像分割，可以较好地提取脑部图像。(The region growing method can extract brain image well.)

说明：  matlab工具箱，用于构建和重建可分离的正交可操纵金字塔。(a matlab toolbox for building and reconstructing separable quadrature steerable pyramids.)

利用粒子滤波方法识别和跟踪动态目标，英文文献，经典(use the method of particle filter to track moving objects)

小波域稀疏表示，正交匹配追踪算法压缩感知重建(wavelet transform Compressive sensing)

经典的跟踪算法，可以对人脸进行实时的跟踪。(Classical tracking algorithm can be real-time face tracking.)