PDlib - A PHP extension for Dlib

Requirements

Dlib 19.13+
PHP 7.0+
C++14
libx11-dev (on Ubuntu: sudo apt-get install libx11-dev)

Dependencies

Dlib

Install Dlib as shared library

git clone https://github.com/davisking/dlib.git
cd dlib/dlib
mkdir build
cd build
cmake -DBUILD_SHARED_LIBS=ON ..
make
sudo make install

Installation

git clone https://github.com/goodspb/pdlib.git
cd pdlib
phpize
./configure --enable-debug
# you may need to indicate the dlib install location
# PKG_CONFIG_PATH=/usr/local/lib/pkgconfig ./configure --enable-debug
make
sudo make install

Configure PHP installation

vim youpath/php.ini

Append the content below into php.ini

[pdlib]
extension="pdlib.so"

Tests

For tests, you will need to have bz2 extension installed. On Ubuntu, it boils to:

sudo apt-get install php-bz2

After you successfully compiled everything, just run:

make test

Usage

General Usage

Good starting point can be tests/integration_face_recognition.phpt. Check that first.

Basically, if you just quickly want to get from your image to 128D descriptor of faces in image, here is really minimal example how:

<?php
$img_path = "image.jpg";
$fd = new CnnFaceDetection("detection_cnn_model.dat");
$detected_faces = $fd->detect($img_path);
foreach($detected_faces as $detected_face) {
  $fld = new FaceLandmarkDetection("landmark_model.dat");
  $landmarks = $fld->detect($img_path, $detected_face);
  $fr = new FaceRecognition("recognition_model.dat");
  $descriptor = $fr->computeDescriptor($img_path, $landmarks);
  // Optionally use descriptor later in `dlib_chinese_whispers` function
}

Location from where to get these models can be found on DLib website, as well as in tests/integration_face_recognition.phpt test.

Specific use cases

face detection

If you want to use HOG based approach:

<?php

// face detection
detected_faces = dlib_face_detection("image.jpg");
// $detected_faces is indexed array, where values are assoc arrays with "top", "bottom", "left" and "right" values

If you want to use CNN approach (and CNN model):

<?php
$fd = new CnnFaceDetection("detection_cnn_model.dat");
$detected_faces = $fd->detect("image.jpg");
// $detected_face is indexed array, where values are assoc arrays with "top", "bottom", "left" and "right" values

CNN model can get you slightly better results, but is much, much more demanding (CPU and memory, GPU is also preferred).

face landmark detection

<?php

// face landmark detection
$landmarks = dlib_face_landmark_detection("~/a.jpg");
var_dump($landmarks);

Additionally, you can also use class-based approach:

$rect = array("left"=>value, "top"=>value, "right"=>value, "bottom"=>value);
// You can download a trained facial shape predictor from:
// http://dlib.net/files/shape_predictor_5_face_landmarks.dat.bz2
$fld = new FaceLandmarkDetection("path/to/shape/predictor/model");
$parts = $fld->detect("path/to/image.jpg", $rect);
// $parts is integer array where keys are associative values with "x" and "y" for keys

Note that, if you use class-based approach, you need to feed bounding box rectangle with values obtained from dlib_face_detection. If you use dlib_face_landmark_detection, everything is already done for you (and you are using HOG face detection model).

face recognition (aka getting face descriptor)

<?php

$fr = new FaceRecognition($model_path);
$landmarks = array(
    "rect" => $rect_of_faces_obtained_with_CnnFaceDetection,
    "parts" => $parts_obtained_with_FaceLandmarkDetection);
$descriptor = $fr->computeDescriptor($img_path, $landmarks);
// $descriptor is 128D array

chinese whispers

Provides raw access to dlib's chinese_whispers function. Client need to build and provide edges. Edges are provided as numeric array. Each element of this array should also be numeric array with 2 elements of long type.

Returned value is also numeric array, containing obtained labels.

<?php
// This example will cluster nodes 0 and 1, but would leave 2 out.
// $labels will look like [0,0,1].
$edges = [[0,0], [0,1], [1,1], [2,2]];
$labels = dlib_chinese_whispers($edges);

luzfcb / pdlib