tensorflow test

利用小波变换进行图像去噪处理，包括各种软硬阈值的选取函数(Using Wavelet Transform for Image Denoising, including a variety of select hard and soft threshold function)

首先将RGB转换到HSV，再根据阈值进行HSV图像的分割(First convert RGB to HSV, then HSV image segmentation based on threshold)

说明：  sift 图像特征提取算法源码，对研究sift算法很有帮助。(The source code of sift image feature extraction algorithm is very helpful to the research of sift algorithm.)

一种改进后的海森矩阵算法代码，针对医学图像处理效果比较好，可以识别较细的血管。(An improved Hessian matrix algorithm code, medical image processing effects for better, you can identify smaller vessels.)

this is a program for image compression

说明：  这是一个用C++语言实现一个指纹识别程序，要求满足FVC国际竞赛标准，并在学校建立的指纹库上实测。程序里应用到的技术有：中值滤波、直方图均衡化、脊线方向提取、Gabor滤波、指纹细化、特征提取、特征点过滤、基于局部特征点的特征匹配(This is a C++ Language realization of a fingerprint identification procedure, the requirements FVC meet international competition standards, and established in the school library on the measured fingerprint. Program applied to the technology are as follows: median filter, histogram equalization, the direction of the ridge line extraction, Gabor filtering, fingerprint thinning, feature extraction, feature point filtering, based on local feature points of the feature matching)

提出了一种基于非下采样Contourlet 变换的自适应图像去噪方法 首先对噪声图像进行非下采样Contourlet 变换，得到各 个尺度各个方向子带的系数，再根据该系数的能量自适应地调整去噪阈值 实验表明，与Contourlet 多尺度阈值去噪 Contourlet 自 适应阈值去噪相比，该方法在保留图像边缘细节的同时，能提高图像的PSNR 值，减少了Gibbs 现象(A new adaptive method of image de-noising based on NonSubsampled Contourlet Transform （NSCT）is presented.Firstly， the noisy image is decomposed into a set of multiscale and multidirectional frequency subbands by NSCT， and then the threshold is adapted automatically according to the energy of subband coefficients.Compared with the multi-scale threshold using contourlet transform and using adaptive threshold based on contourlet transform， the simulation results show that the performance of this method is superior in PSNR， meanwhile it can maintain the image edge and reduce the Gibbs phenomen)

说明：  直方图匹配，或者叫做直方图规定化。在matlab下实现，但是没有使用IPT中的histeq进行规定化。(Histogram matching,or histogram specification is the basic method in the digit image processing. We provide a implementation in matlab without using IPT function histeq)

基于大气模型进行图像的去雾处理，以及均衡化的去雾处理(mage removal based on atmospheric model)

1．用Matlab产生正弦波,矩形波,以及白噪声信号，并显示各自时域波形图 2．进行FFT变换，显示各自频谱图，其中采样率，频率、数据长度自选 3．做出上述三种信号的均方根图谱,功率图谱,以及对数均方根图谱 4．用IFFT傅立叶反变换恢复信号，并显示恢复的正弦信号时域波形图( 1. Using Matlab generated sine wave, square wave, and the white noise signal, and displays the respective time-domain waveform 2. The FFT transform, showing each spectrum, where the sampling rate, frequency, data length choice 3. Rms signal to make the three maps, power spectra, as well as the number of rms spectra 4. With the inverse Fourier transform IFFT restore the signal and displays the recovered time-domain waveform sinusoidal signal)