PHP and OpenCV libraries: How to implement face recognition-AI-php.cn

Face recognition is a method that uses computer vision technology to automatically identify and verify faces. In this article, we will implement face recognition using PHP and OpenCV libraries. We will introduce the installation of the OpenCV library, the installation of the PHP extension, and the detailed steps for face recognition using PHP and the OpenCV library.

PHP and OpenCV libraries: How to implement face recognition

Preparation

Before we start, we need to install some necessary software and extensions.

1. Install the OpenCV library

First, we need to install the OpenCV library.Following are the steps to install OpenCV library on Ubuntu system:

Open a terminal and run the following command to install the necessary dependencies:

sudo apt-get update
sudo apt-get install build-essential cmake
sudo apt-get install libgtk2.0-dev libjpeg-dev libtiff5-dev libjasper-dev libopenexr-dev libavcodec-dev libavformat-dev libswscale-dev libv4l-dev libxvidcore-dev libx264-dev libatlas-base-dev gfortran python2.7-dev python3.6-dev

Copy after login

Download OpenCV source Code and unzip it:

wget -O opencv.zip https://github.com/opencv/opencv/archive/4.2.0.zip
unzip opencv.zip

Copy after login

Create a directory for compilation:

mkdir buildcd build

Copy after login

Use CMake to generate the compiled file:

cmake -D CMAKE_BUILD_TYPE=RELEASE -D CMAKE_INSTALL_PREFIX=/usr/local ..

Copy after login

Compile and install OpenCV:

make -j4
sudo make install
sudo ldconfig

Copy after login

2. Install PHP extension

After installing the OpenCV library, we need to install the PHP extension for use in PHP OpenCV library.以下是在Ubuntu系统上安装PHP扩展的步骤：

安装PHP开发工具：

sudo apt-get install php7.4-dev

Copy after login

安装PECL工具：

sudo apt-get install php-pear

Copy after login

使用PECL安装OpenCV扩展：

sudo pecl install opencv

Copy after login

在php.ini文件中添加OpenCV扩展：

extension=opencv.so

Copy after login

重启PHP：

sudo service apache2 restart

Copy after login

人脸检测

在安装和配置好所有必要的软件和扩展后，我们可以开始使用PHP和OpenCV进行人脸识别。以下是人脸检测的步骤：

1、创建一个PHP文件（例如index.php）。

2、导入OpenCV命名空间：

use \OpenCV\{
  CvImage,
  CvVideoCapture,
  CvWindow,
  CvFont,
  CvScalar};

Copy after login

3、创建一个函数来进行人脸检测：

function detectFace($inputImage){
  // 加载人脸识别器和级联分类器  $faceCascade = new \OpenCV\CascadeClassifier();
  $faceCascade->load('/path/to/haarcascade_frontalface_default.xml');

  // 加载输入图像  $image = new CvImage();
  $image->load($inputImage);

  // 将图像转换为灰度图像  $grayImage = $image->clone();
  $grayImage->toGray();

  // 检测人脸  $faces = $faceCascade->detectMultiScale($grayImage);

  // 在图像中绘制人脸矩形框  $rectColor = new CvScalar(0, 255, 0);
  foreach ($faces as $face) {
    $image->rectangle(
      $face->getX(),
      $face->getY(),
      $face->getX() + $face->getWidth(),
      $face->getY() + $face->getHeight(),
      $rectColor,
      2    );
  }

  // 在新窗口中显示图像  $window = new CvWindow('Face Detection');
  $window->showImage($image);
  $window->waitKey();
}

Copy after login

4、调用detectFace函数并传入要检测的图像路径：

detectFace('/path/to/input/image.jpg');

Copy after login

人脸识别

在人脸检测的基础上，我们可以进一步实现人脸识别。以下是实现人脸识别的步骤：

1、创建一个包含人脸图像和类别标签的训练数据集。每个人脸图像都应该有一个唯一的标签。

2、训练一个基于人脸数据集的人脸识别模型。以下是一个训练模型的示例：

function trainModel($trainingDataPath){
  // 加载训练数据集  $images = [];
  $labels = [];
  $labelsMap = [];

  foreach (glob($trainingDataPath . '/*') as $dir) {
    $label = basename($dir);
    $labelsMap[] = $label;

    foreach (glob($dir . '/*.jpg') as $file) {
      $images[] = CvImage::load($file);
      $labels[] = count($labelsMap) - 1;
    }
  }

  // 创建人脸识别模型  $model = new \OpenCV\Face\FisherFaceRecognizer();
  $model->train($images, $labels);

  // 保存模型  $model->save('/path/to/fisher_model.yml');

  // 保存标签映射  file_put_contents('/path/to/labels.txt', implode(PHP_EOL, $labelsMap));
}

Copy after login

3、创建一个函数来进行人脸识别：

function recognizeFace($inputImage){
  // 加载人脸识别模型  $model = new \OpenCV\Face\FisherFaceRecognizer();
  $model->load('/path/to/fisher_model.yml');

  // 加载标签映射  $labelsMap = file('/path/to/labels.txt', FILE_IGNORE_NEW_LINES);

  // 加载输入图像  $image = CvImage::load($inputImage);
  $grayImage = $image->clone();
  $grayImage->toGray();

  // 检测人脸  $faceCascade = new \OpenCV\CascadeClassifier();
  $faceCascade->load('/path/to/haarcascade_frontalface_default.xml');
  $faces = $faceCascade->detectMultiScale($grayImage);

  // 遍历所有检测到的人脸  foreach ($faces as $face) {
    // 提取人脸区域    $faceImage = $grayImage->getROI($face->getX(), $face->getY(), $face->getWidth(), $face->getHeight());

    // 识别人脸    $predictedLabel = $model->predict($faceImage);

    // 在图像中绘制识别结果    $font = new CvFont();
    $font->scale(1);
    $font->color(new CvScalar(0, 255, 0));
    $image->putText(
      $labelsMap[$predictedLabel],
      new CvPoint($face->getX(), $face->getY() - 10),
      $font
    );
    $image->rectangle(
      $face->getX(),
      $face->getY(),
      $face->getX() + $face->getWidth(),
      $face->getY() + $face->getHeight(),
      new CvScalar(0, 255, 0),
      2    );
  }

  // 在新窗口中显示图像  $window = new CvWindow('Face Recognition');
  $window->showImage($image);
  $window->waitKey();
}

Copy after login

4、调用recognizeFace函数并传入要识别的图像路径：

recognizeFace('/path/to/input/image.jpg');

Copy after login

总结

在本文中，我们介绍了使用PHP和OpenCV库实现人脸识别的详细步骤。我们首先安装了OpenCV库和PHP扩展，然后通过组合OpenCV的人脸检测和人脸识别功能，实现了人脸识别的应用。我们提供了具体的代码示例，并附带了安装教程，以帮助读者快速上手。希望这篇文章能够帮助您了解如何使用PHP和OpenCV库来实现人脸识别。

php Learning Recommendation:

Vue3 Laravel8 Uniapp Novice to Practical Development Tutorial

Vue3 TP6 API Social E-commerce system development teaching

"Workerman TP6 Instant Messaging Chat System" is on sale for a limited time!

The above is the detailed content of PHP and OpenCV libraries: How to implement face recognition. For more information, please follow other related articles on the PHP Chinese website!