=====================face detection

这是用于提取图片的hog特征的，是在VS环境下配置opencv库实现的，希望对你有用(This is the picture of the hog for extracting feature is the VS environment configuration opencv library to achieve, and I hope useful to you)

利用opencv中的stitcher类实现的图像拼接

C# SDK for Face++ 人脸检测与分析 检测给定图片(Image)中的所有人脸(Face)的位置和相应的面部属性 检测模式可以是normal(默认) 或者 oneface 。在oneface模式中，检测器仅找出图片中最大的一张脸。 可以是none或者由逗号分割的属性列表。默认为gender, age, race, smiling。目前支持的属性包括：gender, age, race, smiling, 可以为图片中检测出的每一张Face指定一个不包含非法字符且不超过255字节的字符串作为tag，tag信息可以通过 /info/get_face 查询 检测给定人脸(Face)相应的面部轮廓，五官等关键点的位置，包括25点和83点两种模式。(Face Detection and Analysis Detecting the location and corresponding facial attributes of all faces in a given image The detection mode can be normal (default) or oneface. In oneface mode, the detector finds only the largest face in the picture. It can be none or a comma-separated list of attributes. The default is gender, age, race, smiling. Attributes currently supported include gender, age, race, smiling. Each Face detected in a picture can be tagged with a string that does not contain illegal characters and does not exceed 255 bytes. The tag information can be queried by / Info / get_face. The location of key points such as facial contour and facial features for a given face (Face) is detected, including 25 points and 83 points.)

本人做研究生数学建模时候写的代码，关于动态视频，前景识别检测代码，精确度非常高，获得国家二等奖

区域生长算法的改进方法，适合初步学习，具有多个图形处理，最后标注部分代码(The improved method of region growing algorithm is suitable for preliminary study.)

基于openCV的多角度人脸检测，融合了正脸侧脸haar分类器，具有极强鲁棒性(, The integration of multi-angle face detection based on openCV positive face side faces haar classifier, with strong robustness)

opencv毛星云著， 该资源为一个简易demo，很适合初学者练手，opencv作为一门开源的图像库，本着引用和做学术专研的不二法门，进军人工智能市场，必将愈发发展，同时作为一门智能算法，也是可取的。

基于卡尔曼滤波的车辆检测与跟踪，并对车辆技术跟踪

说明：  通过opencv与visualc++加载一副图像，在c++中读取图像(By loading an image opencv with visualc++, to read images in c++)

在很多动作大片中，开启关键之门的密码都是人脸———通过摄像头扫描脸部特征，符合则自动开启大门，反之警铃大作。依次通过人脸检测、人脸对齐、人脸识别三个模块，才能完成验证。人脸支付存在的一大问题就是识别率不高，完美光线下识别率能达到90%，特殊光影下只有60%。以淘宝为例，每天上亿的交易量，即便是99.99%也是远远不够的。