资源列表
Bag-of-visual-words
- SIFT等局部特征的词袋模型实现。包括K-means聚类，直方图特征的形成，以及KNN分类。-SIFT local features such as word bag model implementation. Including K-means clustering to form histogram features, and KNN classification.
gauss
- 用二维随机数生成二维高斯分布随机表面，可调整均方差、相关长度等参数-Generate a two-dimensional Gaussian random numbers using two-dimensional distribution of random surfaces, adjustable both variance, correlation length and other parameters
MOG
- 利用opencv工具包实现混合高斯背景建模，可完成实时的背景抓取-Opencv toolkits use Gaussian mixture background modeling, to be completed in real-time background crawl
PCAmatlab
- 利用PCA方法对数据集进行降维，能提高识别效率。-Use PCA method to reduce the dimensionality of the data set can improve the recognition efficiency.
filtercode
- 引导滤波器实现的立体匹配，得到深度图，数据量较大，运行时间较长。想要一次全部运行，将runstereomatcher.m中的34行改为for testimage = 1：4-The guided stereotactic matched filter to obtain a depth map, the amount of data is large, the running time is longer. Want to run all at once, the runstereomatche
Wavelet_waterprint_process
- 在图像中加入水印，使用2维离散Daubechies小波变换在图像中加入水印，重构图像，通过检测阈值，将阈值对应水印分离。-Add watermark in the image, using two-dimensional discrete wavelet transform Daubechies watermark in the image, the reconstructed image, through the detection threshold, the threshold corre
canny1125
- 自己编的canny改进算法，里面有详细的注释附带检测的图片。对于目标分割用处很大。-Own series of canny improved algorithm, which has detailed notes with pictures of detection. Proved very useful for object segmentation.
video-segmentation
- 本代码实现的是视频中运动目标的分割,采用了三帧差法进行目标检测。-The code is in the video moving target segmentation, using three difference method for target detection.
AnotherSolution
- 多边形凸凹性的判断的实现以及多边形方向判定-Polygon Polygon implementation and direction of the judge' s judgment punch
Faceracognzie
- 通过分割图像，利用TRAINGDM算法训练BP神经网络，对人脸进行识别-By dividing the image , use TRAINGDM training BP neural network algorithm for face recognition
Test2
- 实现基于OPENCV的摄像头目标识别,笔记本识别出人脸-OPENCV achieve target recognition camera-based notebook to identify the face
code
- 各种图像放大算法的比较和峰值信噪比的源码-Comparison of various image zoom algorithms and peak signal to noise ratio of the source