Harris

用Graph cuts 来解决磁共振图像的 phase unwrapping。(Graph cuts used to solve the magnetic resonance images of the phase unwrapping.)

说明：  使用PCA将数据集降维，降维后将特征脸还原(Use PCA to reduce the dimension of the data set, reduce the feature face after dimension reduction)

图像处理的空域滤波增强部分程序代码的编写(Spatial filtering of the image processing, enhanced part of the program code is written)

在梯度基础上使用加权平均提取指纹方向场然后使用高斯滤波进行平滑处理(Extraction of fingerprint orientation field using weighted average based on gradient)

matlab最小二乘法曲面拟合程序可以得到函数具体解析式(Matlab least squares surface fitting program can get the specific analytic expression of the function)

) 使用分块的主成分分析方法(PCA)对人脸图像进行压缩编码。针对PCA方法计算量大的缺点,首先把问题转化成奇异值分解(SVD)问题,然后设计了特征空间的更新算法,通过递推,简化每一步计算的计算量,达到了实时编码的要求。 4) 在Windows平台下基于Video for Windows(VFW)接口开发了人脸视频图像编码和解码的实验系统,该系统实现了图像采集、图像显示、编码、解码等功能。() The use of sub-blocks of principal component analysis (PCA) on the human face image coding. PCA method for large defects, first of all the issues into a singular value decomposition (SVD) problem, and then design a feature space of the update algorithm, recursive, simplifying the calculation of the amount calculated at each step to achieve real-time encoding requirements. 4) In the Windows platform based on Video for Windows (VFW) interfaces, the human face image coding and decoding video experimental system, the system achieved image acquisition, image display, encoding, decoding functions.)

指纹识别预处理，预处理包括图像增强，图像细化，寻找奇异点，细节检测(Preprocessing of fingerprint identification)

该程序主要实现图像的无损压缩，具体算法包括： （1）整数小波变换＋SPIHT无损压缩 其中整数小波可以采用JPEG2000提供的5－3小波，也可以采用S＋P变换，本程序均已给出。 （2）DPCM＋算术编码 DPCM采用的是JPEG中采用的方法； （3）JPEG－LS标准 JPEG－LS的无损压缩性能非常好，超过JPEG2000。(The program mainly realize lossless image compression, the specific algorithm include: (1) integer wavelet transform SPIHT Integer Wavelet lossless compression which can be used to provide the 5-3 wavelet JPEG2000, can also adopt the S P transform, the procedures are given. (2) DPCM arithmetic coding DPCM is used in JPEG methodology (3) JPEG-LS standard JPEG-LS lossless compression of very good performance, more than JPEG2000.)

对两帧图像进行ICM调试.包含内容：1估计姿态和摄像机内参数；2两幅图做四次金字塔；3利用输入特征点采用ransac算法估计姿态R, T。4sift特征匹配；5摄像机标定。(ICM images on two debugging. Contains: 1 Estimated posture and intrinsic parameters 2 two plans to do four pyramids 3 using the input feature points using ransac algorithm estimates gesture R, T. 4sift feature matching 5 camera calibration.)

图像降噪、贝叶斯估计、盲源降噪,采用盲源方法进行噪声估计，再进行贝叶斯非局部均值降噪(Image denoising, Bias estimation and blind source noise，The blind source method for noise estimation, then the Bayes non local mean noise reduction)