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. (2019-07-22, C#, 4KB, 下载6次)
该资料中包含了《OpenCV By Example》中文版以及例程程序，该书的目录如下所示：
第1章 OpenCV的探险之旅;
第2章 OpenCV基础知识介绍;
第3章 图形用户界面和基本滤波;
第4章 深入研究直方图和滤波器;
第5章 自动光学检测、目标分割和检测;
第6章 学习目标分类;
第7章 识别人脸部分并覆盖面具;
第8章 视频监控、背景建模和形态学操作;
第9章 学习对象跟踪;
第10章 文本识别中的分割算法;
第11章 使用Tesseract识别文本;
The archive contains the Chinese version of "OpenCV By Example" and the example program. The contents of this book are as follows:
Chapter 1 An expedition to OpenCV;
Chapter 2 OpenCV basic knowledge introduction;
Chapter 3 Graphical user interface and basic filtering;
Chapter 4 In-depth study of histogram and filter;
Chapter 5 Automatic optical detection, object segmentation and detection;
Chapter 6 Learning object classification;
Chapter 7 Identify the face part and cover the mask.
Chapter 8 Video surveillance, background modeling and morphological operation.
Chapter 9 Learning object tracking;
Chapter 10 The segmentation algorithm in text recognition;
Chapter 11 Using Tesseract to identify the text; (2018-02-22, C/C++, 64867KB, 下载30次)
基于opencv的手势识别
1、hog : hog特征提取
2、hogTainer ： 样本训练，利用了SVM方式
3、gestureRecognition ： 手势识别，利用了训练得出的线性分类器进行识别分类
训练样本在hog文件夹中，对应gesture_a到gesture_e ，五种手势
Gesture recognition 1, hog: hog feature extraction 2, hogTainer: sample training, the use of SVM mode 3, gestureRecognition: gesture recognition using training resulting linear classifier for recognition and classification of training samples in hog folder corresponding to gesture_a gesture_e, five kinds of gestures (2016-08-12, Visual C++, 13015KB, 下载40次)
利用opencv实现烟雾检测，利用的特征是：在［0.625Hz，2.5Hz］带宽的频率范围内，监
测图像序列的灰度值通过一个二阶的滤波器后，子信号的频率值的变化，观察其峰峰值的个数，当峰值个数超过2次的时候即可判定为烟雾点
Smoke detection using opencv realized, is characterized by the use: In the [0.625Hz, 2.5Hz] bandwidth, frequency range, the gray value of the image sequence monitored by a second-order filter, the sub-signal of the frequency value changes were observed the number of its peak, when the peak number of more than two times when the smoke point can be determined (2013-08-14, Visual C++, 568KB, 下载153次)
OpenCV的全称是：Open Source Computer Vision Library 。OpenCV 拥有包括 300 多个C函数的跨平台的中、高层 API。它不依赖于其它的外部库——尽管也可以使用某些外部库。OpenCV 为Intel® Integrated Performance Primitives (IPP) 提供了透明接口。 这意味着如果有为特定处理器优化的的 IPP 库， OpenCV 将在运行时自动加载这些库
OpenCV stands: Open Source Computer Vision Library. OpenCV has included more than 300 cross-platform C functions, the high-level API. It does not depend on other external libraries- although you can use some external libraries. OpenCV is Intel Integrated Performance Primitives (IPP) provides a transparent interface. This means that if for specific processor optimized for IPP library, OpenCV will automatically load them at run-time libraries (2013-07-16, Visual C++, 13197KB, 下载2次)
本文来讲讲Mat存储的像素该怎么对其进行扫描？扫描的效率又如何？
文章参考opencv自带的tutiol及其code。
实验功能：
可以手动选择打开电脑上的图片。
对4种对Mat矩阵的扫描方法效率进行比较，这4种扫描方法分别为：连续内存直接访问；用迭代器进行访问；直接像素点进行访问；用LUT查找表进行访问；这4种扫描方法可以在菜单栏进行选择。
这些访问完后是将每个访问到的像素点的像素进行压缩，压缩间隔可以在软件的界面中进行修改。
软件的下端显示每种次扫描方法扫描打开的图片所用的平均时间(程序中设定为求100次的平均值)
This article Laijiang Jiang Mat stored in pixels how to scan them? Scan efficiency and how? The article reference the opencv comes tutiol its code. Experimental feature: you can manually choose to open the picture on the computer. Four kinds of scan efficiency matrix Mat these four kinds of scanning methods: continuous direct memory access access iterator direct pixel access access the LUT lookup table these four kinds of scanning methods select in the menu bar. Visit after each access to the pixels of the pixel to be compressed, compression interval can be modified in the software interface. The bottom of the software shows the average time used by each sub-scan method to scan to open the picture (the program is set to seek 100 times the average) (2012-07-22, C/C++, 170KB, 下载53次)
利用opencv和vs2005编写的MFC界面，可利用网络摄像头或者easycap（usb图像采集器）实时动态显示图像，附有截图说明。以前也上传了一个相同的，据反映有点问题，后发现附加依赖项里加错了lib，现已改好，可直接运行。注意opencv要装好，并在程序中设置好相关选项。希望能帮到大家。
Use opencv and vs2005 writing the MFC interface, can use a network of cameras or easycap (usb image collector) real-time dynamic display images, with screenshots show. Upload a before the same, according to reflect a little problem, found that rely on an additional add wrong lib, has changed and can be run directly. Note opencv to pack good, and in the program set up relevant option. The hope can help you. (2012-02-23, Visual C++, 10041KB, 下载156次)
VC6.0 + QT4 + OpenCV 环境。演示视频地址http://v.youku.com/v_show/id_XMzA3MTU2Mzky.html
(低角度车流量检测器)一种基于车头特征的车辆检测，很好的应用于低角度摄像机。与地感线圈车检器相比，解决了跟车问题。与传统基于背景建模的检测方法相比，解决了道路拥堵问题并解决了传统方法受限于摄像机必须架设很高的问题。同时能有效去除自行车、摩托车、行人的干扰，准确率可达95 。 all rights reserved by birdwcp
VC6.0+ QT4+ OpenCV environment. Demo video address http://v.youku.com/v_show/id_XMzA3MTU2Mzky.html
(Low-angle traffic flow detector) based on the characteristics of the front vehicle detection, good used low-angle camera. Vehicle inspection and a sense coil compared to solve the problem with the car. Based on background modeling with traditional detection methods, to solve the problem of road congestion and limited solution to the traditional methods of problem of the high camera must be set up. At the same time can effectively remove the bicycle, motorcycle, pedestrian interference, accuracy rate of up to 95 . all rights reserved by birdwcp (2011-12-27, Visual C++, 28341KB, 下载231次)
本文分析比较了传统运动目标检测的3种主要方法:背景图像差分法、时态差分法和光流法,在此基础上给出了一种背景图像预测算法,大大减少了因为背景变化而产生的目标检测误差。本文基于OpenCV设计出改进的运动目标检测与跟踪算法,实现了运动目标的跟踪,并在VC++编译环境下,利用USB摄像头作为视频采集器,通过观察实验结果可以看出,本文的运动目标检测算法能够正确地检测出视频图像中的运动目标,而且在检测性能上优于普通的自适应背景差分法。
OpenCV-based motion tracking system of some articles I read, I feel you can, share (2011-11-25, Visual C++, 1037KB, 下载132次)
神手遙控 HandRemoteControl v1.4
增加手掌五指輕晃，為KMPlayer播放/暫停指令。
讓你作簡報、看電影免用遙控器, 使用環境： 需搭配PC-Camera 請在亮度足夠環境 使用者距離Camera約2～5公尺，正面對Camera 使用方式： -執行本程式 -使用Power Point簡報時, 以手指均勻連續畫順時針可到下一張，逆時針可到上一張 -使用KMPlayer觀看電影時，以順時針可以快轉，逆時針可以倒轉 需安裝OpenCV
Hand of God to increase remote palm fingers Qinghuang HandRemoteControl v1.4 for KMPlayer play/pause commands. Allows you to make presentations, watching movies with a remote control-free use of the environment: the need with the PC-Camera light enough in the environment around the user from Camera 2 to 5 meters, positive use of the Camera:- the implementation of this program- using Power Point presentation, even with your fingers can be continuously drawn to the next clockwise, counterclockwise to the previous one can be- to use KMPlayer watch a movie, you can quickly turn clockwise, counterclockwise to reverse the need to install OpenCV (2010-12-23, Visual C++, 132KB, 下载7次)