matlab-eyes--detected-code

利用灰度积分投影直接对人脸图像进行检测和眼睛定位是一种常用的算法,但是直接采用该算法会受到背 景、特征等因素的影响,识别准确率较低。提出了一种基于最大类间方差阈值和区域膨胀相结合的检测与定位算法。该算 法首先计算最大类间方差设置阈值,把灰度图像转换为二值图像并检测出人脸区域,然后通过对该人脸区域中的连通区域 进行膨胀及连通性处理,精确定位眼睛坐标。实验表明,此算法可靠,具有较好的识别效果。(face recognized)

包括计算共生矩阵、对共生矩阵归一化、对共生矩阵计算能量、熵、惯性矩、相关4个纹理参数、求能量、熵、惯性矩、相关的均值和标准差作为最终8维纹理特征 (Including the calculation of co-occurrence matrix, normalized to the co-occurrence matrix, co-occurrence matrix calculation on the energy, entropy, moment of inertia, related to four texture parameters, find the energy, entropy, moment of inertia, related to the mean and standard deviation as the final eight-dimensional texture feature)

在做图像处理时用遇到图像的阈值分割，这是几种图像的阈值分割方法(In doing image processing with experience image threshold segmentation, which is one of several image thresholding segmentation method)

给予matlab的图像分割例程，内附相关图片及程序。(Image segmentation routine MATLAB, containing the relevant pictures and program. )

说明：  数字摄影测量中的最小二乘匹配算法，带示例图像(Feature points matching using LSTM)

经典主成分分析法，实现高光谱图像的降维处理 (Classical principal component analysis, to achieve high spectral image dimensionality reduction)

说明：  利用稀疏表示 omp k-svd训练字典的方法，实现了图像去噪(The sparse representation of OMP k-svd training dictionary is used to realize Image Denoising)

说明：  多尺度加权的rentinex算法实现图像增强，里面用到了高斯滤波对图像进行滤波，可以去除大部分的噪声，平滑图像效果较突出，3个高斯尺度对图像做滤波后加权在一起得到多尺度信息的图像。(Multiscale weighted rentinex algorithm is used to enhance the image. Gauss filter is used to filter the image, which can remove most of the noise. The smoothing effect of the image is prominent. Three Gaussian scales are filtered and weighted together to get the multi-scale information of the image.)

对离散特征点进行直线拟合，效果很好，文件中还有平面拟合的例子，以及三焦点张量的部分主要程序，希望对需要的人有帮助(Discrete feature point fitting a straight line, with good results, part of the main program file plane fitting example trifocal tensor hope that those who need help)

说明：  把两幅有相互重叠部分的图片拼接在一起，合成一副大的图片(There is some overlap between the two pictures spliced together, synthesized a large picture)