人脸性别检测库 [英] Face gender detection library

问题描述

我正在寻找可以检测网络摄像头流中的人脸并检测性别的 SDK 或库.Windows 下免费或付费，C++ 或 C#.有什么建议吗?

I'm looking for an SDK or library that can detect faces in a webcam stream, and detect gender. Free or paid, C++ or C# under Windows. Any suggestions?

推荐答案

我用 Fisherfaces 做了一个性别分类实验方法 在我的博客中:http://bytefish.de/blog/gender_classification 得到了几何归一化图像的识别率为 98%.我正在为 GNU Octave/MATLAB、Python 和 C++ 提供 Fisherfaces 方法的实现.

I did an experiment on gender classification with the Fisherfaces method in my blog at: http://bytefish.de/blog/gender_classification and got a 98% recognition rate for geometrically normalized images. I am providing an implementation of the Fisherfaces method for GNU Octave/MATLAB, Python and C++.

libfacerec，一个用于 OpenCV C++ API(BSD 许可)的现代人脸识别库，包括一个实现费雪脸方法.libfacerec 的最新版本位于:

libfacerec, a modern face recognition library for the OpenCV C++ API (BSD license), includes an implementation of the Fisherfaces method. The latest revision of the libfacerec is available at:

• https://github.com/bytefish/libfacerec

libfacerec 是一个 CMake 项目，具有完善的 API，还有一个关于性别分类的教程.您可以在以下位置查看文档的 HTML 版本:

libfacerec comes as a CMake project with a well-documented API, there's also a tutorial on gender classification. You can see a HTML version of the documentation at:

• http://www.bytefish.de/dev/libfacerec/ (性别分类示例)

我在以下位置提供 Python 和 GNU Octave/MATLAB 的 Fisherfaces 实现:

I am providing a Python and GNU Octave/MATLAB implementation of Fisherfaces at:

• http://www.github.com/bytefish/facerec
• http://www.bytefish.de/blog/face_recognition_with_opencv2

对于检测图像中的人脸，只需使用 CascadeClassifier 随 OpenCV 一起提供，但不要期望在未对齐的数据上获得完美的识别结果.

For detecting a face in an image simply use the CascadeClassifier coming with OpenCV, but don't expect perfect recognition results on misaligned data.

上面有一个使用 Eigenfaces 的答案，它不起作用.特征脸使用主成分分析，一种不了解类的无监督算法.如果您将数据嵌入到的子空间可以区分性别，那么您一定很幸运.也许它通过找到最接近的图像和相关的性别来产生一些积极的预测.

There's an answer above to use Eigenfaces, which won't work. Eigenfaces use a Principal Component Analysis, an unsupervised algorithm that doesn't know about classes. You must be really lucky if the subspace you embed your data into discriminates between genders. Perhaps it generates some positive predictions by finding the closest image and associated gender.

这篇关于人脸性别检测库的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！