Face Recognition Attendance System Using Python (With Code)

A D VA NC E D C O M PUT E R VI S I O N IMAGE I M A G E A NA LYS I S PYT HO N

Introduction

In this tutorial, you will learn how to build a face recognition system using computer vision in Python. Face
recognition is an image classification problem and is a step further to face detection. In face detection, we
only detect the location of the human face in an image, but in the face recognition classifier, we make a
system that can identify humans.

“Face recognition is a broad challenge of verifying or identifying people in pictures or videos. Big tech giants are still working to make a
faster and more accurate face recognition model.”

Learning Objectives

Getting an overview of face recognition algorithm in Artificial Intelligence and its real-world
applications.
Learning to preprocess and organize the data for training the face recognition model.
Learning to build the face recognition model using OpenCV and face recognition libraries.
Learning how to integrate the face recognition model with an attendance system.

This article was published as a part of the Data Science Blogathon

Table of Contents

1. Real-World Applications of Face Recognition

2. Traditional Face Recognition Algorithms
3. Deep Learning for Face Recognition
4. How Does a Traditional Face Recognition Model Work?
5. Steps to Build a Face Recognition System
. Building a Face Recognition System
7. Read Webcam for Real-Time Recognition
. Challenges Faced by Face Recognition Systems
9. Conclusion
10. Frequently Asked Questions
Q1. What is Haar cascade algorithm for face detection in Python?
Q2. What is the difference between generic and specific classifiers in Python?
Q3. What is a CNN, and how is it used in facial recognition?

Real-World Applications of Face Recognition

Face recognition is currently being used to make the world safer, smarter, and more convenient.

Some of its most common use cases include finding missing persons, solving retail crime, security
identification, identifying accounts on social media, school attendance systems, and recognizing drivers in
cars.

There are several methods to perform facial recognition depending on the performance and complexity.

Traditional Face Recognition Algorithms

During the 1990s, holistic approaches were used for face recognition. Handcrafted local descriptors
became popular In the early 1920s, and then the local feature learning approaches were followed in the late
2000s. Nowadays, face recognition and face detection algorithms that are widely used and are implemented
in OpenCV are as follows:

Eigenfaces (1991)
Local Binary Patterns Histograms (LBPH) (1996)
Fisherfaces (1997)
Scale Invariant Feature Transform (SIFT) (1999)
Speed Up Robust Features (SURF) (2006)

Each method follows a different approach to extracting and matching the image information with the input
image.

Fischer-faces and Eigenfaces have almost similar approaches as well as SURF and SIFT.

LBPH is a simple yet very efficient method but is slow compared to modern face recognizers.

These algorithms are not faster compared to modern days face-recognition algorithms. Traditional
algorithms can’t be trained only by taking a single picture of a person.

Deep Learning for Face Recognition

Some of the widely used Deep Learning-based face recognition systems are:

DeepFace
DeepID series of systems
VGGFace
FaceNet

Face recognizers generally take face images and find the important points such as the corner of the
mouth, an eyebrow, eyes, nose, lips, etc. The coordinates of these points are called facial feature points.
There are 66 such points. In this way, a different technique for finding feature
points give different results.
source: pinterest

How Does a Traditional Face Recognition Model Work?

1. Face Detection: Face detector algorithms locate faces and draw bounding boxes around faces and
keep the coordinates of bounding boxes.
2. Face Alignments: Normalize the faces to be consistent with the training database.
3. Feature Extraction: Extract features of faces that will be used for training and recognition tasks.
4. Face Recognition: Matching the face against one or more known faces in a prepared database.

In the traditional method of face recognition, we had separate modules to perform these 4 steps, which
was painful. -In this article, you will see a library that combines all these 4 steps in a single step.

Steps to Build a Face Recognition System

Step 1: Install libraries

We need to install 2 libraries in order to implement face recognition.

dlib: Dlib is a modern C++ toolkit containing machine learning algorithms and tools for creating complex
software in C++ to solve real-world problems.

# installing dlib pip install dlib

face recognition: The face_recognition library, created and maintained by Adam Geitgey, wraps around dlib
facial recognition functionality.

# installing face recognition pip install face recognition

Opencv for some image pre-processing.

# installing opencv pip install opencv

Note: if in case you encounter any error while installing dlib, i would recommend you to install the C++ development toolkit using
vs_code community
Step 2: Impor t libraries

Now that you have downloaded all the important libraries, let’s import them to build the system.

import cv2 import numpy as np import face_recognition

Step 3: Load images

After importing libraries you need to load an image.

face_recognition library loads images in the form of BGR, in order to

print the image you should convert it into RGB using OpenCV.

imgelon_bgr = face_recognition.load_image_file('elon.jpg') imgelon_rgb =

cv2.cvtColor(imgelon_bgr,cv2.COLOR_BGR2RGB) cv2.imshow('bgr', imgelon_bgr) cv2.imshow('rgb', imgelon_rgb)

cv2.waitKey(0)

As you see, RGB looks natural, so you will always change the channel to RGB.

Step 4: Find the face location and draw bounding boxes

You need to draw a bounding box around the faces in order to show if the human face has been detected
or not.

imgelon =face_recognition.load_image_file('elon.jpg') imgelon = cv2.cvtColor(imgelon,cv2.COLOR_BGR2RGB) #----

------Finding face Location for drawing bounding boxes------- face =

face_recognition.face_locations(imgelon_rgb)[0] copy = imgelon.copy() #-------------------Drawing the

Rectangle------------------------- cv2.rectangle(copy, (face[3], face[0]),(face[1], face[2]), (255,0,255), 2)

cv2.imshow('copy', copy) cv2.imshow('elon',imgelon) cv2.waitKey(0)

Step 5: Train an image for face recognition

This library is made in such a way that it automatically finds the face and works on only faces, so you don’t
need to crop the face out of pictures.
Training:

At this stage, we convert the train image into some encodings and store the encodings with the given
name of the person for that image.

train_elon_encodings = face_recognition.face_encodings(imgelon)[0]

Testing:

For testing, we load an image and convert it into encodings, and now match encodings with the stored
encodings during training. This matching is based on finding maximum similarity. When you find the
encoding matching the test image, you get the name associated with train encodings.

# lets test an image test = face_recognition.load_image_file('elon_2.jpg') test = cv2.cvtColor(test,

cv2.COLOR_BGR2RGB) test_encode = face_recognition.face_encodings(test)[0]

print(face_recognition.compare_faces([train_encode],test_encode))

face_recognition.compare_faces returns True if the person in both images is the same, otherwise, it
returns False.

Building a Face Recognition System

Step 1: Impor t the necessary libraries.

import cv2 import face_recognition import os import numpy as np from datetime import datetime import pickle

Step 2: Define a folder path where your training image dataset will be stored.

path = 'student_images'

Note: for training, we only need to drop the training images in the path directory and the image name must be
person_name.jpg/jpeg format.

For example: as you see in my student_images path, I have 6 persons. hence our model can recognize only
these 6 persons. you can add more pictures in this directory for more people to be recognized.

Step 3: Now, create a list to store person_name and image array.

Traverse all image files present in the path directory, read images, and append the image array to the image
list and file name to classNames.

images = [] classNames = []mylist = os.listdir(path) for cl in mylist: curImg = cv2.imread(f'{path}/{cl}')

images.append(curImg) classNames.append(os.path.splitext(cl)[0])

Step 4: Create a function to encode all the train images and store them in a variable encoded_face_train.

def findEncodings(images): encodeList = [] for img in images: img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)

encoded_face = face_recognition.face_encodings(img)[0] encodeList.append(encoded_face) return encodeList

encoded_face_train = findEncodings(images)

Step 5: Create a function that will create aAttendance.csv file to store the attendance with time.

Note: here you need to create Attendance.csv file manually and give the path in the function

def markAttendance(name): with open('Attendance.csv','r+') as f: myDataList = f.readlines() nameList = [] for

line in myDataList: entry = line.split(',') nameList.append(entry[0]) if name not in nameList: now =

datetime.now() time = now.strftime('%I:%M:%S:%p') date = now.strftime('%d-%B-%Y') f.writelines(f'n{name},
{time}, {date}')

Step 6: With open(“filename.csv”,’r+’) create a file, and ‘r+’ mode is used to open a file for reading and
writing.

We first check if the name of the attendee is already available in attendance.csv. We won’t write
attendance again. If the attendee’s name is not available in attendance.csv we will write the attendee’s
name with a time of function call.

Read Webcam for Real-Time Recognition

# take pictures from webcam cap = cv2.VideoCapture(0)while True: success, img = cap.read() imgS =
cv2.resize(img, (0,0), None, 0.25,0.25) imgS = cv2.cvtColor(imgS, cv2.COLOR_BGR2RGB) faces_in_frame =
face_recognition.face_locations(imgS) encoded_faces = face_recognition.face_encodings(imgS,

faces_in_frame)for encode_face, faceloc in zip(encoded_faces,faces_in_frame): matches =

face_recognition.compare_faces(encoded_face_train, encode_face) faceDist =
face_recognition.face_distance(encoded_face_train, encode_face) matchIndex = np.argmin(faceDist)

print(matchIndex) if matches[matchIndex]: name = classNames[matchIndex].upper().lower() y1,x2,y2,x1 = faceloc

# since we scaled down by 4 times y1, x2,y2,x1 = y1*4,x2*4,y2*4,x1*4 cv2.rectangle(img,(x1,y1),(x2,y2),
(0,255,0),2) cv2.rectangle(img, (x1,y2-35),(x2,y2), (0,255,0), cv2.FILLED) cv2.putText(img,name, (x1+6,y2-5),

cv2.FONT_HERSHEY_COMPLEX,1,(255,255,255),2) markAttendance(name) cv2.imshow('webcam', img) if cv2.waitKey(1)

& 0xFF == ord('q'): break

Resize the image by 1/4 only for the recognition part. output frame will be of the original size.
Resizing improves the Frame per Second.
face_recognition.face_locations()is called on the resized image(imgS) .for face bounding box
coordinates must be multiplied by 4 in order to overlay on the output frame.
face_recognition.distance() returns an array of the distance of the test image with all images present
in our train directory.
The index of the minimum face distance will be the matching face.
 After finding the matching name, we call the markAttendance function.
Draw a bounding box using cv2.rectangle().
We put the matching name on the output frame using cv2.putText().

Attendance Report

Challenges Faced by Face Recognition Systems

Although building facial recognition seems easy, it is not as easy in real-world images that are being taken
without any constraints. There are several challenges that are faced by the Facial Recognitions System, as
follows:

Illumination: It changes the face’s appearance drastically. It is observed that even slight changes in
lighting conditions cause a significant impact on its results.
Pose: Facial Recognition systems are highly sensitive to the pose, Which may result in faulty
recognition or no recognition if the database is only trained on frontal face view.
Facial Expressions: Different expressions of the same individual are another significant factor that
needs to be taken into account. Modern Recognizers can easily deal with it, though.
Low Resolution: The recognizer must be trained on a good-resolution picture. Otherwise, the model will
fail to extract features.
Aging: With increasing age, the human face features shape, lines, and texture changes which are yet
another challenge.

Conclusion

In this article, we discussed how to create a face recognition system using the face_recognition library
and made an attendance system. The face recognition model can be trained using a dataset of labeled face
images, and the model accuracy can be improved by using techniques such as face alignment and data
augmentation. You can further design GUI using Tkinter or Pyqt for the face recognition attendance
system.

Key Takeaways

Face recognition technology can be used to build practical systems for attendance tracking, security
access control, and more.
The face recognition system can be built using Python programming language and popular libraries
such as OpenCV and face recognition.
Once the face recognition model is built, it can be integrated with an attendance system using Python
programming and database management tools.

Frequently Asked Questions

Q1. What is Haar cascade algorithm for face detection in Python?

A. Haar cascade algorithm is an open-source machine learning-based approach used for object detection
in images, commonly used for face detection in Python. It works by training a classifier using positive and
negative samples of the object of interest (in this case, faces) and then applying the classifier to new
images to detect the object. The algorithm uses a series of features known as Haar features to distinguish
between the object and the background. The algorithm is efficient and can detect faces in real-time video
streams, making it a popular choice for face detection applications.

Q2. What is the difference between generic and specific classifiers in Python?

A. A generic classifier is useful for detecting a wide range of objects but may not be as accurate for
detecting specific objects. A specific classifier, on the other hand, is more accurate for detecting its
designated object but may not be useful for detecting other types of objects.

Q3. What is a CNN, and how is it used in facial recognition?

A. A Convolutional Neural Network (CNN) is a type of neural network commonly used for image
classification and object recognition. In facial recognition, a CNN can be trained on large datasets of facial
images to learn how to identify and recognize different faces.

The media shown in this ar ticle is not owned by Analytics Vidhya and are used at the Author’s discretion

Article Url - https://www.analyticsvidhya.com/blog/2021/11/build-face-recognition-attendance-system-

using-python/

Thetechwriters