Как установить face recognition python на windows

In this article, we will learn how to install Face Recognition in Python on Windows. Recognize and manipulate faces from Python or from the command line with the world’s simplest face recognition library. Built using dlib’s state-of-the-art face recognition built with deep learning.

Installing Face Recognition on Windows :

Prerequisites:

Face Recognition module can only be installed for Python version 3.7 and 3.8.

Step 1: Install git for Windows

Step 2: Clone this repository and go inside the folder using the following commands

git clone https://github.com/RvTechiNNovate/face_recog_dlib_file.git
cd face_recog_dlib_file

Step 3: Enter the following command to install dlib and cmake using pip

Python 3.7:
pip install dlib-19.19.0-cp37-cp37m-win_amd64.whl

Python 3.8:
pip install dlib-19.19.0-cp38-cp38-win_amd64.whl

pip install cmake

Method 1: Using pip to install Face Recognition Package

Follow the below steps to install the Face Recognition package on Windows using pip:

Step 1: Install the latest Python3 in Windows 

Step 2: Check if pip and python are correctly installed.

python --version
pip --version

Step 3: Upgrade your pip to avoid errors during installation.

pip install --upgrade pip

Step 4: Enter the following command to install Face Recognition using pip3.

pip install face-recognition

Method 2: Using setup.py to install Face Recognition 

Follow the below steps to install the Face Recognition on Windows using the setup.py file:

Step 1: Download the latest source package of Face Recognition for python3 from here.

curl https://files.pythonhosted.org/packages/6c/49/75dda409b94841f01cbbc34114c9b67ec618265084e4d12d37ab838f4fd3/face_recognition-1.3.0.tar.gz > face_recognition-1.3.0.tar.gz

Step 2: Extract the downloaded package using the following command.

tar -xzvf face_recognition-1.3.0.tar.gz

Step 3: Go inside the folder and Enter the following command to install the package.

cd face_recognition-1.3.0
python setup.py install

Verifying Face Recognition installation on Windows :

Make the following import in your python terminal to verify if the installation has been done properly:

import face_recognition

If there is any error while importing the module then is not installed properly.

Переводчик Елена Борноволокова специально для Нетологии адаптировала статью Файзана Шайха о том, как создать модель распознавания лиц и в каких сферах ее можно применять.

Введение

За последние годы компьютерное зрение набрало популярность и выделилось в отдельное направление. Разработчики создают новые приложения, которыми пользуются по всему миру.
В этом направлении меня привлекает концепция открытого исходного кода. Даже технологические гиганты готовы делиться новыми открытиями и инновациями со всеми, чтобы технологии не оставались привилегией богатых.

Одна из таких технологий — распознавание лиц. При правильном и этичном использовании эта технология может применяться во многих сферах жизни.

В этой статье я покажу вам, как создать эффективный алгоритм распознавания лиц, используя инструменты с открытым исходным кодом. Прежде чем перейти к этой информации, хочу, чтобы вы подготовились и испытали вдохновение, посмотрев это видео:

Распознавание лиц: потенциальные сферы применения

Приведу несколько потенциальных сфер применения технологии распознавания лиц.

Распознавание лиц в соцсетях. Facebook заменил присвоение тегов изображениям вручную на автоматически генерируемые предложения тегов для каждого изображения, загружаемого на платформу. Facebook использует простой алгоритм распознавания лиц для анализа пикселей на изображении и сравнения его с соответствующими пользователями.

Распознавание лиц в сфере безопасности. Простой пример использования технологии распознавания лиц для защиты личных данных — разблокировка смартфона «по лицу». Такую технологию можно внедрить и в пропускную систему: человек смотрит в камеру, а она определяет разрешить ему войти или нет.

Распознавание лиц для подсчета количества людей. Технологию распознавания лиц можно использовать при подсчете количества людей, посещающих какое-либо мероприятие (например, конференцию или концерт). Вместо того чтобы вручную подсчитывать участников, мы устанавливаем камеру, которая может захватывать изображения лиц участников и выдавать общее количество посетителей. Это поможет автоматизировать процесс и сэкономить время.

Настройка системы: требования к аппаратному и программному обеспечению

Рассмотрим, как мы можем использовать технологию распознавания лиц, обратившись к доступным нам инструментам с открытым исходным кодом.

Я использовал следующие инструменты, которые рекомендую вам:

• Веб-камера (Logitech C920) для построения модели распознавания лиц в реальном времени на ноутбуке Lenovo E470 ThinkPad (Core i5 7th Gen). Вы также можете использовать встроенную камеру своего ноутбука или видеокамеру с любой подходящей системой для анализа видео в режиме реального времени вместо тех, которые использовал я.
• Предпочтительно использовать графический процессор для более быстрой обработки видео.
• Мы использовали операционную систему Ubuntu 18.04 со всем необходимым ПО.

Прежде чем приступить к построению нашей модели распознавания лиц, разберем эти пункты более подробно.

Шаг 1: Настройка аппаратного обеспечения

Проверьте, правильно ли настроена камера. С Ubuntu это сделать просто: посмотрите, опознано ли устройство операционной системой. Для этого выполните следующие шаги:

1. Прежде чем подключить веб-камеру к ноутбуку, проверьте все подключенные видео устройства, напечатав в командной строке ls /dev/video*. В результате выйдет список всех видео устройств, подключенных к системе.
2. Подключите веб-камеру и задайте команду снова. Если веб-камера подключена правильно, новое устройство будет отражено в результате выполнения команды.
3. Также вы можете использовать ПО веб-камеры для проверки ее корректной работы. В Ubuntu для этого можно использовать программу «Сheese».

Шаг 2: Настройка программного обеспечения

Шаг 2.1: Установка Python

Код, указанный в данной статье, написан с использованием Python (версия 3.5). Для установки Python рекомендую использовать Anaconda – популярный дистрибутив Python для обработки и анализа данных.

Шаг 2.2: Установка OpenCV

OpenCV  – библиотека с открытым кодом, которая предназначена для создания приложений компьютерного зрения. Установка OpenCV производится с помощью pip:

pip3 install opencv-python

Шаг 2.3: Установите face_recognition API

Мы будем использовать face_recognition API, который считается самым простым API для распознавания лиц на Python во всем мире. Для установки используйте:

pip install dlib
pip install face_recognition

Внедрение

После настройки системы переходим к внедрению. Для начала, мы создадим программу, а затем объясним, что сделали.

Пошаговое руководство

Создайте файл face_detector.py и затем скопируйте приведенный ниже код:

# import libraries
import cv2
import face_recognition
 
# Get a reference to webcam
video_capture = cv2.VideoCapture("/dev/video1")
 
# Initialize variables
face_locations = []
 
while True:
	
# Grab a single frame of video
	
ret, frame = video_capture.read()
 
	
# Convert the image from BGR color (which OpenCV uses) to RGB color (which face_recognition uses)
	
rgb_frame = frame[:, :, ::-1]
 
	
# Find all the faces in the current frame of video
	
face_locations = face_recognition.face_locations(rgb_frame)
 
	
# Display the results
	
for top, right, bottom, left in face_locations:
    	
# Draw a box around the face
    	
cv2.rectangle(frame, (left, top), (right, bottom), (0, 0, 255), 2)
 
	
# Display the resulting image
	
cv2.imshow('Video', frame)
 
	
# Hit 'q' on the keyboard to quit!
	
if cv2.waitKey(1) & 0xFF == ord('q'):
    	
break
 
# Release handle to the webcam
video_capture.release()
cv2.destroyAllWindows()

Затем запустите этот файл Python, напечатав:

python face_detector.py

Если все работает правильно, откроется новое окно с запущенным режимом распознавания лиц в реальном времени.

Подведем итоги и объясним, что сделал наш код:

1. Сначала мы указали аппаратное обеспечение, на котором будет производиться анализ видео.
2. Далее сделали захват видео в реальном времени кадр за кадром.
3. Затем обработали каждый кадр и извлекли местонахождение всех лиц на изображении.
4. В итоге, воспроизвели эти кадры в форме видео вместе с указанием на то, где расположены лица.

Пример применения технологии распознавания лиц

На этом все самое интересное не заканчивается. Мы сделаем еще одну классную вещь: создадим полноценный пример применения на основе кода, приведенного выше. Внесем небольшие изменения в код, и все будет готово.

Предположим, что вы хотите создать автоматизированную систему с использованием видеокамеры для отслеживания, где спикер находится в данный момент времени. В зависимости от его положения, система поворачивает камеру так, что спикер всегда остается в центре кадра.
Первый шаг — создайте систему, которая идентифицирует человека или людей на видео и фокусируется на местонахождении спикера.

Разберем, как это сделать. В качестве примера я выбрал видео на YouTube с выступлением спикеров конференции «DataHack Summit 2017».

Сначала импортируем необходимые библиотеки:

import cv2
import face_recognition

Затем считываем видео и устанавливаем длину:

input_movie = cv2.VideoCapture("sample_video.mp4")
length = int(input_movie.get(cv2.CAP_PROP_FRAME_COUNT))

После этого создаем файл вывода с необходимым разрешением и скоростью передачи кадров, аналогичной той, что была в файле ввода.

Загружаем изображение спикера в качестве образца для распознания его на видео:

image = face_recognition.load_image_file("sample_image.jpeg")
face_encoding = face_recognition.face_encodings(image)[0]
 
known_faces = [
face_encoding,
]

Закончив, запускаем цикл, который будет:

• Извлекать кадр из видео.
• Находить все лица и идентифицировать их.
• Создавать новое видео, которое будет сочетать в себе оригинал кадра с указанием местонахождения лица спикера с подписью.

Посмотрим на код, который будет это выполнять:

# Initialize variables
face_locations = []
face_encodings = []
face_names = []
frame_number = 0
 
while True:
	
# Grab a single frame of video
	
ret, frame = input_movie.read()
	
frame_number += 1
 
	
# Quit when the input video file ends
	
if not ret:
    	
break
 
	
# Convert the image from BGR color (which OpenCV uses) to RGB color (which face_recognition uses)
	
rgb_frame = frame[:, :, ::-1]
 
	
# Find all the faces and face encodings in the current frame of video
	
face_locations = face_recognition.face_locations(rgb_frame, model="cnn")
	
face_encodings = face_recognition.face_encodings(rgb_frame, face_locations)
 
	
face_names = []
	
for face_encoding in face_encodings:
    	
# See if the face is a match for the known face(s)
    	
match = face_recognition.compare_faces(known_faces, face_encoding, tolerance=0.50)
 
    	
name = None
    	
if match[0]:
        	
name = "Phani Srikant"
 
    	
face_names.append(name)
 
	
# Label the results
	
for (top, right, bottom, left), name in zip(face_locations, face_names):
    	
if not name:
        	
continue
 
       # Draw a box around the face
    	
cv2.rectangle(frame, (left, top), (right, bottom), (0, 0, 255), 2)
 
    	
# Draw a label with a name below the face
    	
cv2.rectangle(frame, (left, bottom - 25), (right, bottom), (0, 0, 255), cv2.FILLED)
        font = cv2.FONT_HERSHEY_DUPLEX
    	
cv2.putText(frame, name, (left + 6, bottom - 6), font, 0.5, (255, 255, 255), 1)
 
	
# Write the resulting image to the output video file
	
print("Writing frame {} / {}".format(frame_number, length))
	
output_movie.write(frame)
 
# All done!
input_movie.release()
cv2.destroyAllWindows()

Код даст вам вот такой результат:

От редакции

Курсы «Нетологии» по теме:

• онлайн-профессия «Python-разработчик»
• онлайн-профессия «Data Scientist»

The Goal

In this tutorial, I’ll go over some example usages of the Python face_recognition library to:

• Detect faces in images
• Detect facial features on a detected face (like eyebrows and nose)
• Check for matches of detected faces

All images and code snippets are provided on this post along with step-by-step instructions and explanations as to what is going on. This tutorial is aimed towards Windows 10 but Linux and macOS users will most likely find this easier as they can skip some of the prerequisites.

Here are some relevant links for face_recognition if you need/want them:

• Documentation: face-recognition.readthedocs.io
• Source code: github.com/ageitgey/face_recognition
• PyPI: pypi.org/project/face-recognition

Installing The «face_recognition» Library

Prerequisites (Windows)

To install the face_recognition library on Windows you will need the following installed:

• CMake
• Visual Studio C++ build tools

If you do not have these, you will get errors like,

CMake must be installed to build the following extensions: _dlib_pybind11

Which is telling you that CMake isn’t installed, or,

You must use Visual Studio to build a python extension on windows.  If you
are getting this error it means you have not installed Visual C++.  Note
that there are many flavours of Visual Studio, like Visual Studio for C#
development.  You need to install Visual Studio for C++.

Which is telling you that you need Visual Studio C++ build tools.

CMake

To install CMake, go to cmake.org/download/ and download the appropriate installer for your machine. I am using 64bit Windows 10 so I will get cmake-<version>-win64-x64.msi. After downloading the setup file, install it.

While installing CMake, add CMake to the system PATH environment variable for all users or the current user so it can be found easily.

After the installation is complete, open a terminal and execute cmake. This should show the usage for CMake. If it did not, make sure you selected the option to add it to the PATH environment variable.

You will need to close and re-open your terminal/application for the PATH variable to update so the cmake binary can be identified.

Visual Studio C++ Build Tools

Unlike Linux, C++ compilers for Windows are not included by default in the OS. If we visit the WindowsCompilers page on the Python wiki we can see there is information on getting a standalone version of Visual C++ 14.2 compiler without the need for Visual Studio. If we visit the link from that wiki section, we will be brought to a Microsoft download page. On this download page, you will want to download «Build Tools for Visual Studio 2019».

When vs_buildtools__<some other stuff>.exe has downloaded, run the exe and allow it to install a few things before we get to the screen below. When you get to this screen, make the selections I have:

After clicking «Install», wait for the installation to complete and restart.

Now that CMake and the required build tools are installed, we can then continue to installing the face_recognition library.

Installing face_recognition and Verifying The Installation

To install the face_recognition library, execute the following in a terminal:

python -m pip install face-recognition

This might take a bit longer to install than usual as dlib needs to be built using the tools installed from above. To validate that the library was installed successfully, try to import the library in Python using the following:

No errors should be raised.

Installing PIL

For this tutorial, we will also be using Pillow / PIL which will help us crop and draw on images. This is not required to use the face_recognition library but will be required in this tutorial to prove/show results. To install it, execute the following:

python -m pip install Pillow

To validate the library was installed, try to import PIL in Python using the following:

No errors should be raised.

Detecting A Face In An Image

Now that we have set the face_recognition library and PIL up, we can now start detecting faces. To start off, we will detect a face in an image with just one person.

single-person.jpg

First, we want to import face_recognition and some helpers from PIL.

import face_recognition
from PIL import Image, ImageDraw

Now load your image using face_recognition.load_image_file.

image = face_recognition.load_image_file('single-person.jpg')

image now contains our image in a format that face_recognition can detect faces with. To identify the location of the face in this image, call face_recognition.face_locations and pass the image.

face_locations = face_recognition.face_locations(image)

face_locations will now contain a list of face locations. Each face location is a tuple of pixel positions for (top, right, bottom, left) — we need to remember this for when we use it later.

Since there is only one face in this image, we would expect there to be only one item in this list. To check how many faces were detected, we can get the length of the list.

amount = len(face_locations)
print(f'There are {amount} face locations')

For the example image I have provided above and am using for this tutorial, this has told me there was one face detected.

To get the location of this face to use later, we can then just get the first element out of the list.

first_face_location = face_locations[0]

To see what’s in this, you can call:

print(first_face_location)

Which will print out something like:

These are the pixel positions for (top, right, bottom, left) which we will use to create a box and crop with soon.

Identifying The Detected Face

To identify where the face is that was detected in the image, we will draw a red box on the bounds that were returned by face_recognition.face_locations.

First, we need to create a PIL image from the image that was loaded using face_recognition.load_image_file. Doing this will allow us to use features offered by PIL.

img = Image.fromarray(image, 'RGB')

Now that we have the PIL image, we need to create an object to help us draw on the image. Before we do this, we will also copy the image into a new object so that when we crop the face out later there won’t be a red box still around it.

img_with_red_box = img.copy()
img_with_red_box_draw = ImageDraw.Draw(img_with_red_box)

Now that we have an object to help us draw on the image, we will draw a rectangle using the dimensions returned earlier.

To draw a box, we need two points, the top left and the bottom right as x and y coordinates. Since we got back (top, right, bottom, left), we need to make these (left, top), (right, bottom); the basic translation can be seen below.

img_with_red_box_draw.rectangle(
    [
        (first_face_location[3], first_face_location[0]),
        (first_face_location[1], first_face_location[2])
    ],
    outline="red",
    width=3
)

We needed (left, top), (right, bottom) to get (x, y) (x, y) points.

In that step we have also set outline="red" to make the box red and width=3 to make the box 3 pixels wide.

To see the final result, we call:

This will open the image in the default image viewer. The image should look like this:

Cropping Out The Detected Face

Aside from drawing a box, we can also crop out the face into another image.

Using the original image that we didn’t draw on (because we drew on the copied image), we can call img.crop providing the dimensions from before.

img_cropped = img.crop((
    first_face_location[3],  # Left x
    first_face_location[0],  # Top y
    first_face_location[1],  # Right x
    first_face_location[2]   # Bottom y
))

img.crop returns a copy of the original image being copped so you do not need to copy before-hand if you want to do something else with the original image.

img_cropped now contains a new cropped image, to display it, we can call .show() again.

Final Code For This Section

import face_recognition
from PIL import Image, ImageDraw

# Detecting the faces
image = face_recognition.load_image_file('single-person.jpg')  # Load the image
face_locations = face_recognition.face_locations(image)  # Detect the face locations
first_face_location = face_locations[0]  # Get the first face

# Convert the face_recognition image to a PIL image
img = Image.fromarray(image, 'RGB')

# Creating the image with red box
img_with_red_box = img.copy()  # Create a copy of the original image so there is not red box in the cropped image later
img_with_red_box_draw = ImageDraw.Draw(img_with_red_box)  # Create an image to draw with
img_with_red_box_draw.rectangle(  # Draw the rectangle on the image
    [
        (first_face_location[3], first_face_location[0]),  # (left, top)
        (first_face_location[1], first_face_location[2])  # (right, bottom)
    ],
    outline="red",  # Make the box red
    width=3  # Make the box 3px in thickness
)
img_with_red_box.show()  # Open the image in the default image viewer

# Creating the cropped image
img_cropped = img.crop((  # Crop the original image
    first_face_location[3],
    first_face_location[0],
    first_face_location[1],
    first_face_location[2]
))
img_cropped.show()  # Open the image in the default image viewer

Detecting Multiple Faces In An Image

We saw before that face_recognition.face_locations returns an array of tuples corresponding with the locations of faces. This means we can use the same methods above, but loop over the result of face_recognition.face_locations when drawing and cropping.

I will use the following as my image. It has 5 faces visible, 2 of which are slightly blurred.

group-of-people.jpg

Once again, we want to import face_recognition and some helpers from PIL.

import face_recognition
from PIL import Image, ImageDraw

Then load the new image using face_recognition.load_image_file and detect the faces using the same methods as before.

image = face_recognition.load_image_file('group-of-people.jpg')
face_locations = face_recognition.face_locations(image)

If we print out face_locations (print(face_locations)), we can see that 5 faces have been detected.

[
    (511, 1096, 666, 941),
    (526, 368, 655, 239)
    (283, 1262, 390, 1154),
    (168, 1744, 297, 1615),
    (271, 390, 378, 282)
]

Since we now have more than one face, taking the first one does not make sense — we should loop over them all and perform our operations in the loop.

Before we continue, we should also create the PIL image we will be working with.

img = Image.fromarray(image, 'RGB')

Identifying The Detected Faces

Like before, we need to copy the original image for later (optional) and create an object to help us draw.

img_with_red_box = img.copy()
img_with_red_box_draw = ImageDraw.Draw(img_with_red_box)

Now we can loop all the faces and create rectangles.

for face_location in face_locations:
    img_with_red_box_draw.rectangle(
        [
            (face_location[3], face_location[0]),
            (face_location[1], face_location[2])
        ],
        outline="red",
        width=3
    )

And once again, look at the image:

Not bad eh!

Cropping Out The Detected Faces

Just like drawing many boxes, we can also crop all the detected faces. Use a for-loop again, crop the image in each loop and then showing the image.

for face_location in face_locations:
    img_cropped = img.crop((face_location[3], face_location[0], face_location[1], face_location[2]))
    img_cropped.show()

You will have many images open on your machine in separate windows; here they are all together:

I have downscaled these images for the tutorial but yours will be cropped at whatever resolution you put in.

Final Code For This Section

import face_recognition
from PIL import Image, ImageDraw

# Load image and detect faces
image = face_recognition.load_image_file("group-of-people.jpg")
face_locations = face_recognition.face_locations(image)

# Create the PIL image to copy and crop
img = Image.fromarray(image, 'RGB')

img_with_red_box = img.copy()  # Make a single copy for all the red boxes
img_with_red_box_draw = ImageDraw.Draw(img_with_red_box)  # Get our drawing object again
for face_location in face_locations:  # Loop over all the faces detected this time
    img_with_red_box_draw.rectangle(  # Draw a rectangle for the current face
        [
            (face_location[3], face_location[0]),
            (face_location[1], face_location[2])
        ],
        outline="red",
        width=3
    )
img_with_red_box.show()  # Open the image in the default image viewer

for face_location in face_locations:  # Loop over all the faces detected
    img_cropped = img.crop((  # Crop the current image like we did last time
        face_location[3],
        face_location[0],
        face_location[1],
        face_location[2]
    ))
    img_cropped.show()  # Show the image for the current iteration

Identifying Facial Features

face_recognition also has a function face_recognition.face_landmarks which works like face_recognition.face_locations but will return a list of dictionaries containing face feature positions rather than the positions of the detected faces themselves.

Going back to the image with one person in it, we can import everything again, load the image and call face_recognition.face_landmarks.

import face_recognition
from PIL import Image, ImageDraw

image = face_recognition.load_image_file('single-person.jpg')
face_landmarks_list = face_recognition.face_landmarks(image)  # The new call

Now if we print face_landmarks_list, the object will look a bit different.

[
    {
        'chin': [(315, 223), (318, 248), (321, 273), (326, 296), (335, 319), (350, 339), (370, 354), (392, 365), (415, 367), (436, 363), (455, 351), (469, 336), (479, 318), (486, 296), (488, 273), (490, 251), (489, 229)],
        'left_eyebrow': [(329, 194), (341, 183), (358, 180), (375, 182), (391, 189)],
        'right_eyebrow': [(434, 189), (448, 184), (461, 182), (474, 184), (483, 194)],
        'nose_bridge': [(411, 209), (411, 223), (412, 238), (412, 253)],
        'nose_tip': [(394, 269), (403, 272), (412, 275), (421, 272), (428, 269)],
        'left_eye': [(349, 215), (360, 208), (373, 207), (384, 216), (372, 218), (359, 219)],
        'right_eye': [(436, 216), (446, 208), (458, 208), (467, 216), (459, 219), (447, 219)],
        'top_lip': [(374, 309), (388, 300), (402, 296), (411, 298), (420, 296), (434, 301), (448, 308), (442, 308), (420, 307), (411, 308), (402, 307), (380, 309)],
        'bottom_lip': [(448, 308), (434, 317), (421, 321), (411, 322), (401, 321), (388, 317), (374, 309), (380, 309), (402, 309), (411, 310), (421, 309), (442, 308)]
    }
]

There is quite a bit of stuff here. For each facial feature (i.e. chin, left eyebrow, right eyebrow, etc), a corresponding list contains x and y tuples — these x and y points are x and y coordinates on the image.

PIL offers a .line() method on the drawing object we have been uses which takes a list of x and y points which is perfect for this situation. To start we will need a drawing object.

img = Image.fromarray(image, 'RGB')  # Make a PIL image from the loaded image
img_draw = ImageDraw.Draw(img)  # Create the draw object

In this example we will not copy the image as this is the only time we will be using it

Now that we have the object to help us draw, plot all these lines using the first dictionary in the list returned above.

face_landmarks = face_landmarks_list[0]  # Get the first dictionary of features
img_draw.line(face_landmarks['chin'])
img_draw.line(face_landmarks['left_eyebrow'])
img_draw.line(face_landmarks['right_eyebrow'])
img_draw.line(face_landmarks['nose_bridge'])
img_draw.line(face_landmarks['nose_tip'])
img_draw.line(face_landmarks['left_eye'])
img_draw.line(face_landmarks['right_eye'])
img_draw.line(face_landmarks['top_lip'])
img_draw.line(face_landmarks['bottom_lip'])

Now we can call .show() on our image to look at it:

Final Code For This Section

import face_recognition
from PIL import Image, ImageDraw

# Load the image and detect face landmarks for each face within
image = face_recognition.load_image_file('single-person.jpg')
face_landmarks_list = face_recognition.face_landmarks(image)

# Make a PIL image from the loaded image and then get a drawing object
img = Image.fromarray(image, 'RGB')
img_draw = ImageDraw.Draw(img)

# Draw all the features for the first face
face_landmarks = face_landmarks_list[0]  # Get the first object corresponding to the first face
img_draw.line(face_landmarks['chin'])
img_draw.line(face_landmarks['left_eyebrow'])
img_draw.line(face_landmarks['right_eyebrow'])
img_draw.line(face_landmarks['nose_bridge'])
img_draw.line(face_landmarks['nose_tip'])
img_draw.line(face_landmarks['left_eye'])
img_draw.line(face_landmarks['right_eye'])
img_draw.line(face_landmarks['top_lip'])
img_draw.line(face_landmarks['bottom_lip'])

img_with_face_landmarks.show()  # Show the image for the current iteration

Matching Detected Faces

The face_recognition library also provides the function face_recognition.compare_faces which can be used to compare detected faces to see if they match.

There are two arguments for this function which we will use:

• known_face_encodings: A list of known face encodings
• face_encoding_to_check: A single face encoding to compare against the list

For this section, we will be getting 2 face encodings for the same person and checking if they are in another image.

Here are the two images with a known face in each:

elon-musk-1.jpg
elon-musk-2.png

And here are the images we will check to see if Elon is in them:

elon-musk-in-group.jpg
group-of-people.jpg

Getting Face Encodings

To get known face encodings, we can use face_recognition.face_encodings. This function takes an image that contains a face and the locations of faces in images to use.

Typically you would only use one location of a face in a single image to create one encoding but if you have multiple of the same face in one image you can provide more than one location.

The process of what we need to do to get our known face encodings is:

1. Load in the first image
2. Detect faces in the image to get the face locations
3. Verify there is only one face and select the first face
4. Call face_recognition.face_encodings with the image and the one face location
5. Repeat 1 through 5 for the second image

We have done steps 1-3 previously, so we can do it here again:

import face_recognition

# Load image and detect faces
image = face_recognition.load_image_file("elon-musk-1.jpg")
face_locations = face_recognition.face_locations(image)

And to validate one face was detected:

print(len(face_locations))  # Should be 1

Now we can call face_recognition.face_encodings and provide the image and found location.

face_location = face_locations[0]  # We only want an encoding for the first face. There may be more than one face in images you use so I am leaving this here as a note.
face_encodings = face_recognition.face_encodings(image, [face_location])

The parameter known_face_locations is optional and face_recognition will detect all the faces in the image automatically when getting face encodings if known_face_locations is not supplied. For this part, I am demonstrating it this way to validate there is only one detected face in the image.

Since we specified an array of known_face_locations, we know that there will be only one encoding returned so we can take the first.

elon_musk_knwon_face_encoding_1 = face_encodings[0]

We can now repeat this process for the other image

image = face_recognition.load_image_file("elon-musk-2.png")  # Load the image
face_locations = face_recognition.face_locations(image)  # Get face locations
face_location = face_locations[0]  # Only use the first detected face

face_encodings = face_recognition.face_encodings(image, [face_location])  # Get all face encodings
elon_musk_knwon_face_encoding_2 = face_encodings[0]  # Pull out the one returned face encoding

Now that we have elon_musk_knwon_face_encoding_1 and elon_musk_knwon_face_encoding_2, we can see if they exist in our other two images.

Checking For Matches

To check for matches in an image with face_recognition.compare_faces, we need known face encodings (which we have just gotten above) and a single face encoding to check.

Since we are working with groups of images, we will have to loop the detected faces — although this will also work the same with only one person in the image.

First, we need to get all face decodings out of our first image to compare. I had noted above that the second parameter, known_face_locations, to face_recognition.face_encodings was optional, leaving it out will detect all faces in the image automatically and return face encodings for every face in the image; this is exactly what we want and will remove the intermediate step of detecting faces.

image = face_recognition.load_image_file("elon-musk-in-group.jpg")  # Load the image we are comparing
unknwon_face_encodings = face_recognition.face_encodings(image)  # Get face encodings for everyone in the image

Now that we have our unknown face encodings for all the faces in the group image, we can loop over them and check if either match:

for unknwon_face_encoding in unknwon_face_encodings:
    matches = face_recognition.compare_faces(
        [elon_musk_knwon_face_encoding_1, elon_musk_knwon_face_encoding_2],  # The known face encodings (can be only 1 - less is faster)
        unknwon_face_encoding  # The single unknown face encoding
    )
    print(matches)

If you run this, you will see something like:

[True, True]
[False, False]
[False, False]
[False, False]

Each line is one comparison (stored in matches) and each boolean corresponds to a known face encoding and whether it matched the unknown face encoding.

From the values above, we can see that the first unknown face in the group image matched both known face encodings and the other three unknown faces didn’t match either encoding. Using more than once face encoding can allow you to calculate a better probability but will cost you in speed.

If you run this on the other image, everything returns false.

Final Code For This Section

import face_recognition


# Load elon-musk-1.jpg and detect faces
image = face_recognition.load_image_file("elon-musk-1.jpg")
face_locations = face_recognition.face_locations(image)

# Get the single face encoding out of elon-musk-1.jpg
face_location = face_locations[0]  # Only use the first detected face
face_encodings = face_recognition.face_encodings(image, [face_location])
elon_musk_knwon_face_encoding_1 = face_encodings[0]  # Pull out the one returned face encoding


# Load elon-musk-2.jpg and detect faces
image = face_recognition.load_image_file("elon-musk-2.png")
face_locations = face_recognition.face_locations(image)

# Get the single face encoding out of elon-musk-2.jpg
face_location = face_locations[0]
face_encodings = face_recognition.face_encodings(image, [face_location])
elon_musk_knwon_face_encoding_2 = face_encodings[0]


# Load the image with unknown to compare
image = face_recognition.load_image_file("elon-musk-in-group.jpg")  # Load the image we are comparing
unknwon_face_encodings = face_recognition.face_encodings(image)

# Loop over each unknwon face encoding to see if the face matches either known encodings
print('Matches for elon-musk-in-group.jpg')
for unknwon_face_encoding in unknwon_face_encodings:
    matches = face_recognition.compare_faces(
        [elon_musk_knwon_face_encoding_1, elon_musk_knwon_face_encoding_2],  # The known face encodings (can be only 1 - less is faster)
        unknwon_face_encoding  # The single unknown face encoding
    )
    print(matches)


# Load the other image with unknown to compare
image = face_recognition.load_image_file("group-of-people.jpg")  # Load the image we are comparing
unknwon_face_encodings = face_recognition.face_encodings(image)

# Loop over each unknwon face encoding to see if the face matches either known encodings
print('Matches for group-of-people.jpg')
for unknwon_face_encoding in unknwon_face_encodings:
    matches = face_recognition.compare_faces(
        [elon_musk_knwon_face_encoding_1, elon_musk_knwon_face_encoding_2],
        unknwon_face_encoding
    )
    print(matches)

Additional PIL Help

This tutorial doesn’t focus on PIL but one function you may find useful is img.save(); this saves a file. An example of its usages is img.save('my_image.png') to save a PIL image to my_image.png.

You can find more on PIL in it’s docs and there is plenty of other help online as it is a very old library.

В этой статье мы узнаем, как установить распознавание лиц в Python в Windows. Распознавайте лица и управляйте ими из Python или из командной строки с помощью самой простой в мире библиотеки распознавания лиц. Создан с использованием современного распознавания лиц dlib, созданного с помощью глубокого обучения.

Предпосылки:

Модуль распознавания лиц можно установить только для версий Python 3.7 и 3.8.

Шаг 1. Установите git для Windows

Шаг 2: Клонируйте этот репозиторий и зайдите в папку, используя следующие команды.

git clone https://github.com/RvTechiNNovate/face_recog_dlib_file.git
cd face_recog_dlib_file

Шаг 3: Введите следующую команду, чтобы установить dlib и cmake с помощью pip

Python 3.7:
pip install dlib-19.19.0-cp37-cp37m-win_amd64.whl

Python 3.8:
pip install dlib-19.19.0-cp38-cp38-win_amd64.whl

pip install cmake

Способ 1: использование pip для установки пакета распознавания лиц

Выполните следующие шаги, чтобы установить пакет распознавания лиц в Windows с помощью pip:

Шаг 1. Установите последнюю версию Python3 в Windows.

Шаг 2: Проверьте правильность установки pip и python.

python --version
pip --version

Шаг 3: Обновите ваш pip, чтобы избежать ошибок во время установки.

pip install --upgrade pip

Шаг 4: Введите следующую команду, чтобы установить распознавание лиц с помощью pip3.

pip install face-recognition

Способ 2: использование setup.py для установки распознавания лиц

Выполните следующие шаги, чтобы установить распознавание лиц в Windows с помощью файла setup.py:

Шаг 1: Загрузите последний исходный пакет распознавания лиц для python3 отсюда.

curl https://files.pythonhosted.org/packages/6c/49/75dda409b94841f01cbbc34114c9b67ec618265084e4d12d37ab838f4fd3/face_recognition-1.3.0.tar.gz > face_recognition-1.3.0.tar.gz

Шаг 2: Извлеките загруженный пакет с помощью следующей команды.

tar -xzvf face_recognition-1.3.0.tar.gz

Шаг 3: Войдите в папку и введите следующую команду, чтобы установить пакет.

cd face_recognition-1.3.0
python setup.py install

Проверка установки распознавания лиц в Windows:

Сделайте следующий импорт в вашем терминале Python, чтобы убедиться, что установка была выполнена правильно:

import face_recognition

Если при импорте модуля возникает какая-либо ошибка, значит, он не установлен должным образом.

i have installed the cmake but still dlib is not installing which is required for the installation of face_recognition module

the below mentioned error i am getting whenever i try to install the dlib by using the pip install dlib

ERROR: Complete output from command 'c:userssunilappdatalocalprogramspythonpython37python.exe' -u -c 'import setuptools, tokenize;__file__='"'"'C:\Users\sunil\AppData\Local\Temp\pip-install-oufh_gcl\dlib\setup.py'"'"';f=getattr(tokenize, '"'"'open'"'"', open)(__file__);code=f.read().replace('"'"'rn'"'"', '"'"'n'"'"');f.close();exec(compile(code, __file__, '"'"'exec'"'"'))' bdist_wheel -d 'C:UserssunilAppDataLocalTemppip-wheel-2fd_0qt9' --python-tag cp37:
  ERROR: running bdist_wheel
  running build
  running build_py
  package init file 'dlib__init__.py' not found (or not a regular file)
  running build_ext
  Building extension for Python 3.7.0 (v3.7.0:1bf9cc5093, Jun 27 2018, 04:59:51) [MSC v.1914 64 bit (AMD64)]
  Invoking CMake setup: 'cmake C:UserssunilAppDataLocalTemppip-install-oufh_gcldlibtoolspython -DCMAKE_LIBRARY_OUTPUT_DIRECTORY=C:UserssunilAppDataLocalTemppip-install-oufh_gcldlibbuildlib.win-amd64-3.7 -DPYTHON_EXECUTABLE=c:userssunilappdatalocalprogramspythonpython37python.exe -DCMAKE_LIBRARY_OUTPUT_DIRECTORY_RELEASE=C:UserssunilAppDataLocalTemppip-install-oufh_gcldlibbuildlib.win-amd64-3.7 -A x64'
  -- Building for: NMake Makefiles
  CMake Error in CMakeLists.txt:
    Generator

      NMake Makefiles

    does not support platform specification, but platform

      x64

    was specified.


  CMake Error: CMAKE_C_COMPILER not set, after EnableLanguage
  CMake Error: CMAKE_CXX_COMPILER not set, after EnableLanguage
  -- Configuring incomplete, errors occurred!
  See also "C:/Users/sunil/AppData/Local/Temp/pip-install-oufh_gcl/dlib/build/temp.win-amd64-3.7/Release/CMakeFiles/CMakeOutput.log".
  Traceback (most recent call last):
    File "<string>", line 1, in <module>
    File "C:UserssunilAppDataLocalTemppip-install-oufh_gcldlibsetup.py", line 261, in <module>
      'Topic :: Software Development',
    File "c:userssunilappdatalocalprogramspythonpython37libsite-packagessetuptools__init__.py", line 129, in setup
      return distutils.core.setup(**attrs)
    File "c:userssunilappdatalocalprogramspythonpython37libdistutilscore.py", line 148, in setup
      dist.run_commands()
    File "c:userssunilappdatalocalprogramspythonpython37libdistutilsdist.py", line 966, in run_commands
      self.run_command(cmd)
    File "c:userssunilappdatalocalprogramspythonpython37libdistutilsdist.py", line 985, in run_command
      cmd_obj.run()
    File "c:userssunilappdatalocalprogramspythonpython37libsite-packageswheelbdist_wheel.py", line 192, in run
      self.run_command('build')
    File "c:userssunilappdatalocalprogramspythonpython37libdistutilscmd.py", line 313, in run_command
      self.distribution.run_command(command)
    File "c:userssunilappdatalocalprogramspythonpython37libdistutilsdist.py", line 985, in run_command
      cmd_obj.run()
    File "c:userssunilappdatalocalprogramspythonpython37libdistutilscommandbuild.py", line 135, in run
      self.run_command(cmd_name)
    File "c:userssunilappdatalocalprogramspythonpython37libdistutilscmd.py", line 313, in run_command
      self.distribution.run_command(command)
    File "c:userssunilappdatalocalprogramspythonpython37libdistutilsdist.py", line 985, in run_command
      cmd_obj.run()
    File "C:UserssunilAppDataLocalTemppip-install-oufh_gcldlibsetup.py", line 135, in run
      self.build_extension(ext)
    File "C:UserssunilAppDataLocalTemppip-install-oufh_gcldlibsetup.py", line 172, in build_extension
      subprocess.check_call(cmake_setup, cwd=build_folder)
    File "c:userssunilappdatalocalprogramspythonpython37libsubprocess.py", line 328, in check_call
      raise CalledProcessError(retcode, cmd)
  subprocess.CalledProcessError: Command '['cmake', 'C:\Users\sunil\AppData\Local\Temp\pip-install-oufh_gcl\dlib\tools\python', '-DCMAKE_LIBRARY_OUTPUT_DIRECTORY=C:\Users\sunil\AppData\Local\Temp\pip-install-oufh_gcl\dlib\build\lib.win-amd64-3.7', '-DPYTHON_EXECUTABLE=c:\users\sunil\appdata\local\programs\python\python37\python.exe', '-DCMAKE_LIBRARY_OUTPUT_DIRECTORY_RELEASE=C:\Users\sunil\AppData\Local\Temp\pip-install-oufh_gcl\dlib\build\lib.win-amd64-3.7', '-A', 'x64']' returned non-zero exit status 1.
  ----------------------------------------
  ERROR: Failed building wheel for dlib
  Running setup.py clean for dlib
Failed to build dlib
Installing collected packages: dlib
  Running setup.py install for dlib ... error
    ERROR: Complete output from command 'c:userssunilappdatalocalprogramspythonpython37python.exe' -u -c 'import setuptools, tokenize;__file__='"'"'C:\Users\sunil\AppData\Local\Temp\pip-install-oufh_gcl\dlib\setup.py'"'"';f=getattr(tokenize, '"'"'open'"'"', open)(__file__);code=f.read().replace('"'"'rn'"'"', '"'"'n'"'"');f.close();exec(compile(code, __file__, '"'"'exec'"'"'))' install --record 'C:UserssunilAppDataLocalTemppip-record-89jcoq15install-record.txt' --single-version-externally-managed --compile:
    ERROR: running install
    running build
    running build_py
    package init file 'dlib__init__.py' not found (or not a regular file)
    running build_ext
    Building extension for Python 3.7.0 (v3.7.0:1bf9cc5093, Jun 27 2018, 04:59:51) [MSC v.1914 64 bit (AMD64)]
    Invoking CMake setup: 'cmake C:UserssunilAppDataLocalTemppip-install-oufh_gcldlibtoolspython -DCMAKE_LIBRARY_OUTPUT_DIRECTORY=C:UserssunilAppDataLocalTemppip-install-oufh_gcldlibbuildlib.win-amd64-3.7 -DPYTHON_EXECUTABLE=c:userssunilappdatalocalprogramspythonpython37python.exe -DCMAKE_LIBRARY_OUTPUT_DIRECTORY_RELEASE=C:UserssunilAppDataLocalTemppip-install-oufh_gcldlibbuildlib.win-amd64-3.7 -A x64'
    -- Building for: NMake Makefiles
    CMake Error in CMakeLists.txt:
      Generator

        NMake Makefiles

      does not support platform specification, but platform

        x64

      was specified.


    CMake Error: CMAKE_C_COMPILER not set, after EnableLanguage
    CMake Error: CMAKE_CXX_COMPILER not set, after EnableLanguage
    -- Configuring incomplete, errors occurred!
    See also "C:/Users/sunil/AppData/Local/Temp/pip-install-oufh_gcl/dlib/build/temp.win-amd64-3.7/Release/CMakeFiles/CMakeOutput.log".
    Traceback (most recent call last):
      File "<string>", line 1, in <module>
      File "C:UserssunilAppDataLocalTemppip-install-oufh_gcldlibsetup.py", line 261, in <module>
        'Topic :: Software Development',
      File "c:userssunilappdatalocalprogramspythonpython37libsite-packagessetuptools__init__.py", line 129, in setup
        return distutils.core.setup(**attrs)
      File "c:userssunilappdatalocalprogramspythonpython37libdistutilscore.py", line 148, in setup
        dist.run_commands()
      File "c:userssunilappdatalocalprogramspythonpython37libdistutilsdist.py", line 966, in run_commands
        self.run_command(cmd)
      File "c:userssunilappdatalocalprogramspythonpython37libdistutilsdist.py", line 985, in run_command
        cmd_obj.run()
      File "c:userssunilappdatalocalprogramspythonpython37libsite-packagessetuptoolscommandinstall.py", line 61, in run
        return orig.install.run(self)
      File "c:userssunilappdatalocalprogramspythonpython37libdistutilscommandinstall.py", line 545, in run
        self.run_command('build')
      File "c:userssunilappdatalocalprogramspythonpython37libdistutilscmd.py", line 313, in run_command
        self.distribution.run_command(command)
      File "c:userssunilappdatalocalprogramspythonpython37libdistutilsdist.py", line 985, in run_command
        cmd_obj.run()
      File "c:userssunilappdatalocalprogramspythonpython37libdistutilscommandbuild.py", line 135, in run
        self.run_command(cmd_name)
      File "c:userssunilappdatalocalprogramspythonpython37libdistutilscmd.py", line 313, in run_command
        self.distribution.run_command(command)
      File "c:userssunilappdatalocalprogramspythonpython37libdistutilsdist.py", line 985, in run_command
        cmd_obj.run()
      File "C:UserssunilAppDataLocalTemppip-install-oufh_gcldlibsetup.py", line 135, in run
        self.build_extension(ext)
      File "C:UserssunilAppDataLocalTemppip-install-oufh_gcldlibsetup.py", line 172, in build_extension
        subprocess.check_call(cmake_setup, cwd=build_folder)
      File "c:userssunilappdatalocalprogramspythonpython37libsubprocess.py", line 328, in check_call
        raise CalledProcessError(retcode, cmd)
    subprocess.CalledProcessError: Command '['cmake', 'C:\Users\sunil\AppData\Local\Temp\pip-install-oufh_gcl\dlib\tools\python', '-DCMAKE_LIBRARY_OUTPUT_DIRECTORY=C:\Users\sunil\AppData\Local\Temp\pip-install-oufh_gcl\dlib\build\lib.win-amd64-3.7', '-DPYTHON_EXECUTABLE=c:\users\sunil\appdata\local\programs\python\python37\python.exe', '-DCMAKE_LIBRARY_OUTPUT_DIRECTORY_RELEASE=C:\Users\sunil\AppData\Local\Temp\pip-install-oufh_gcl\dlib\build\lib.win-amd64-3.7', '-A', 'x64']' returned non-zero exit status 1.
    ----------------------------------------
ERROR: Command "'c:userssunilappdatalocalprogramspythonpython37python.exe' -u -c 'import setuptools, tokenize;__file__='"'"'C:\Users\sunil\AppData\Local\Temp\pip-install-oufh_gcl\dlib\setup.py'"'"';f=getattr(tokenize, '"'"'open'"'"', open)(__file__);code=f.read().replace('"'"'rn'"'"', '"'"'n'"'"');f.close();exec(compile(code, __file__, '"'"'exec'"'"'))' install --record 'C:UserssunilAppDataLocalTemppip-record-89jcoq15install-record.txt' --single-version-externally-managed --compile" failed with error code 1 in C:UserssunilAppDataLocalTemppip-install-oufh_gcldlib

can anyone tell me the easiest way to install the face_recognition module for my windows 10