Attendance System Using Face Recognition For Educational Institutes, Essay.

Challenges faced by institutes in keeping handwritten attendance records

Many institutes like schools, colleges, and universities always face challenges of keeping all the handwritten records of the student’s attendants in classes for every class/ batch. Keeping the proxy attendance manually is very tedious and may not be very accurate. For this technique, every student will have to carry their RFID cards for their attendance records daily. An old way of undertaking this process is through calling of names or signing documents which always takes lots of time and is very insecure. Each instruction and school require a very reliable technique that they can employ in the tracking of the student`s attendance. Thus to help reduce this hard and tedious work, a technology of attendance system using face recognition is suggested which makes the use of the pictures of the faces of the students.  This technique of recognizing the student’s faces or a system of verifying the faces of the student from their IDs/video frames and digital pictures becomes more accurate.  

 For the face recognition system to operate effectively, it has to be able to distinguish between the various faces which are stored in the database which are based on the specified information of the faces.  During the attendance, the face gets checked and the images are transferred through a Bluetooth system which is employed for proxy attendance.  This system can be referred to as biometric artificial intelligence (AI) because it has the ability to the identification of students through checking for the patterns in their face forms and features. The use of facial recognition is a technique of tracking attendance which is the fastest and the highest efficient way of handling the records of attendances in such institutions.  In comparison to other available techniques which could have been used as well including fingerprints and voice recognition, but face recognition is the best and the quickest technique. 

To execute a fingerprint -attendance system there must be a mobile device for fingerprint it will have to do this through the use of the RFID.  And the system to have some records, every student will have RFID cards to the card reader.  Despite the fact that the system of face recognition attendance is becoming more common, it is suggested that it should be developed through the use of a trained artificial neuron network (ANN).  The process of this system will involve face detection in the first stages of the operation of the whole system.  After the image (face) has been detected then it will be recognized.  The use of this technique offers various advantages which will be thoroughly analyzed in the paper, but a few advantages can be highlighted like, the system is able to capture images from a relatively far distance of 0.8 m

The aim of the project is to develop an attendance system using the face recognition technique.

To meet the above aim of the project, the following objectives must be met;

• To show a system that helps in detection and recognition of human faces in real-time in various institutions like in colleges and universities to help mark the student`s attendance
• To develop a system that is automated to help improve the operation of older techniques like calling of names and marking of attendance list manually in papers.

Face and Bio-Metric Based Attendance and Security System using RFID and Arduino

  This project is all about the implementation of an RFID system of attendance which is integrated with face recognition for the students in learning institutions but it can also be employed in other fields like in face recognition of employees.  With the integration of biometrics (fingerprint authentication) in the system security of a place can be highly enhanced.  Some audio welcoming messages on the registration of the employee`s attendance can be introduced.  The face recognition system can be coded through the use of Arduino Uno and the codes can be given as below. Arduino can be employed for detection and tracking of faces of students is given in the appendix

Efficiency and reliability of face recognition system for attendance tracking

The Arduino NANO microcontroller is employed in the coding and two servo motors (one for up and down and the other one for right and left) can be used for tracking the face;

 

Figure 1: Showing the circuit diagram for Arduino for face recognition and tracking (https://circuitdigest.com/microcontroller-projects/arduino-face-tracking-robot) 

And the output of this operation can be illustrated in the following diagram;

 

Figure 2: Showing the output of face recognition and tracking using Arduino (https://circuitdigest.com/microcontroller-projects/arduino-face-tracking-robot)  

           The project focuses on the attendance system improvement in learning institutes like colleges and schools. Since there are various bottlenecks of taking manual records of student attendance, such as the cost, inaccuracy, and fake attendance. Thus the face recognition and biometric technology are employed as it helps in solving these bottlenecks.  Traditionally, the face recognition employed was not as accurate as the ones used nowadays. The system of attendance using the face recognition technique was very powerful. In this concept, the images are captured through the closed-circuit television (CCTV) camera in colleges and schools for the porpoise of attendance. After that the system will detect the human face through the features in the face like the nose, the eyes, hair, the mouth as well as various.

           There are various techniques that can be employed for face detection like the use of LBP, Ada-Boast, SNOW, and the SQMT.  After face detection, the methods of face recognition techniques are employed like the HOG (Histogram of oriented Gradient). After this, it will compare the image captured with the stored image in the database. But if the stored image does not match with the captured image then it will store such image under the unknown person database. The outcome of the detection and image recognition can be done using MATLAB codes. The following are the MATLAB codes for face recognition; 

% Create a cascade detector object.

faceDetector = vision.CascadeObjectDetector(); 

% Read a video frame and run the detector.

videoFileReader = vision.VideoFileReader('visionface.avi');

videoFrame      = step(videoFileReader);

bbox            = step(faceDetector, videoFrame);

% Draw the returned bounding box around the detected face.

videoOut = insertObjectAnnotation(videoFrame,'rectangle',bbox,'Face');

figure, imshow(videoOut), title ('Detected face');

 The block diagram below illustrates the working and the process of system of attendance which is based on face recognition through the;

• The features of Haar cascade
• HOG 

 

Figure 3: Showing a block diagram for the working and the process of system of attendance which is based on face recognition (https://www.ijrte.org/wp-content/uploads/papers/v9i1/A2644059120.pdf) 

          The basics behind face recognition is actually the processing of the images.  There are 2 key types of image processing;

• Digital processing: Digital processing entails manipulation of the digital image content through the use of a computer or a laptop, this category also has some two subcategories named:

1. Automated Attendance system
2. Manual Attendance system 

• Analog processing: This is a technology that employs hard copies in manipulation like the use of printouts and photographs.

           There are lots of bottlenecks in dealing with manual attendance such as marking the absent/ present by pen daily which makes it tedious and maintaining all the marked papers which are also very cumbersome.  But all these problems are solved through the use of an automated attendance system. There is also some system that could be proposed but they are affected by serious challenges like the following;

• Bluetooth system: Bluetooth system is not scaled and it also needs at least eight connections at a time.
• Biometric-based system: This system scans the unique part of the human body like a fingerprint to mark attendance but its main challenge is that it takes a lot of time thus it is time-consuming.   
• The use of the RFID system:  This system operates perfectly by swiping the card to the card reader but the challenge comes when the card is lost.

  The key concept in this project is to improve the system attendance using the facial recognition technique. This will minimize the attendance proxy as well improve the system`s accuracy;

Integration of biometrics (fingerprint authentication) for enhancing security

PIR Sensor:  The sensor of Passive Infrared is employed for taking the measurements of the radiation from the object as well as the object`s motion.

Microcontroller: In this case, Arduino UNO/ Arduino NANO is employed as a controller for the implementation of the operation together with the use of the sensors.  

In this system proposed for face, recognition is conducted through the use of sensors and Arduino UNO/ NANO microcontroller. The whole concept is put in various steps as below;

• In the first step, the students will have to fill their registration form together with all their details that are stored in the college database. The student’s picture and the image is also stored in the database; this step is only needed once.   
• The camera will be set at the class entrance together with sensors and an Arduino microcontroller. In this situation the PIR sensor is employed for radiation measurements, the sensor also detects the objects in motions.
• While the students enter into their various classes, the first PIR sensor measures the object`s motion and radiation of the object. And in case the radiation belongs to a human being then the camera will get activated thus it will capture the picture.
• After clicking on the student`s image, the system will have to compare the captured image and compare it with what is in the database, and if the image matches then it gets updated. If not matched to what is given in the database, it will be marked absent. 

The proposed design of the system has some benefits which makes it suitable and highly reliable to be used for attendance system using face recognition technique. Some of these benefits are discussed below;

• This proposed system is capable of handling huge database and the store huge numbers of images for training.
• This system has a higher accurate algorithm that have been employed in many cases as compared to other algorithms.
• Attendance system using face recognition technique system can capture pictures from a distance of 0.8m distance accurately.
• The connectivity of the network is not needed thus no issues such as network problem while the system is under operation.
• The machine interface and Human direct interface is less therefore minimizing lots of errors which will help in increasing the accuracy to the required extent.
• The consumed time for the dataset creation as well as the image training is very less
• This system is very simple to operate with higher accuracy for recording attendance of students in class.
• The speed of the image capturing is higher good as it also work without getting struck 

  The proposed system of attendance is in four various parts where through the use of the Webcam will follow to capture the image of the face.  The diagram below illustrates the picture database of a student for the recognition of face and the record of attendance. A laptop or a computer having an inbuilt web camera is used in this system. For the real-time images, these pictures need to be used in the creation of the database of the students for the recognition of the student`s faces.  For the verification of the presence of the students in a class, a laptop will have to snap the images in a real-time video feed for the face of the student. Then the deep learning of the neural network will be used in the determination of whether the images of the student`s face match anyone in the database or not.  And in case it matches, the system will then determine how far or how close it matches what is in the database then it will identify the students by their name thus making an attendance record.  We have various techniques in which this data can be used for example the data can be used through Microsoft excel.

 

Figure 4: Showing a suggested system of face recognition (https://iopscience.iop.org/article/10.1088/1742-6596/2089/1/012078/pdf)   

 For reasons such as those for security, it is advisable to preserve the faces` images because the identification of our identities is whole dependent on the recognition of our facial features as  the system eliminated the parts of human`s body There should be a method of detecting the area of the face automatically , such that if an image having a face can be generated automatically , then most images of faces should be trimmed for the next recognition of the face, but not every image can be stored.  An efficacy way of realizing this is through the use of cascade classifier known as the Haar feature based system.  This technology is based on machine learning for the negative and positive pictures which are employed in training of the cascade functions.  It is important to note that we must have all the resources for the training of the Haar cascade classifier.  At this point one can see a larger version of face which is already identified in a picture for a sixe of 96x 96.  The scaled image will be stored in the system database or it can be processed in real time through a processor of a real time.  The face detection can be illustrated in the following images;

Arduino Uno as a tool for detection and tracking of faces

 

Figure 5: Showing video screen having a green rectangle for the detected face and on the left is an already cropped face picture for the face recognition (https://iopscience.iop.org/article/10.1088/1742-6596/2089/1/012078/pdf)  

   There are two ways for face recognition and these include either recognition or database based on the image detection and the face focused.  The key component of the algorithms is the deep neural network FaceNet which is employed for the conversion of the facial images to the compact Euclidean space in which there is slight variation in the facial resemblance.  Through the use of FaceNet algorithm, a 128-bit number is generated.  A vector of 128 elements can be generated from the image through dimensional encoding it in a way that the encoding for two images of a person and also a bigger distance separating two images of people for the same person having the same encoding.  Because the FaceNet training takes a huge amount of data and time to process hence we need to load the formerly trained FaceNet to a new inception block of FaceNet which is an already trained model. The resultant model will have a total of 3743280 and every student will have an encoding database of 128. In reduction, we can take 10 images of the faces of the students to help generate 10 encodings. Every student must have his/her own computer for a dictionary object and have their names as keys, their names and coding are kept in the encoded information in the Python language.  And every student is obliged to be present in front of the camera (webcam) when the attendance check is being done.  The accuracy of the face recognition is affected by the settings of the threshold so as to make sure that there is a higher accuracy in the face recognition, the software will fetch a lot of images of one face and identify each face in real-time and produces a result which is based on the collected information which has been received from several accolades.  The flow chart below illustrates a process of face recognition

 

Figure 6: Showing a flow chart for the process of face recognition (https://iopscience.iop.org/article/10.1088/1742-6596/2089/1/012078/pdf)  

The image of the xlsx file given below illustrates the students` recorded attendance list where the first column of the file is filled with the names of the students to create the starting file.  In the attendance record, all the columns are used except the first column for a single class meeting. In case a student has been identified through face recognition for more than 5 times in a total trial of ten, then the student will be considered to be proficient. And this will be recorded in a cell which will be linked to every recognition of student`s row and the class` date is on the column. And because each student who is already identified and has a number is eligible for the college participation, he or she is then regarded to be in class while the student without the number (the record of less than five for the ten trials) is deemed not to be in class;

MATLAB codes for face recognition

Table 1: Showing lists of students with their score out of 10 trials

 

The chart below illustrates the scores of the 14 students for a maximum score of 10 in ten trials of which the students with less than 5 score will be marked as absents while those with above 5 will be marked as present.

Figure 7: Showing a chart of the score of the students

After the system has been used in marking the attendance using face recognition, what will be displayed will only be either present or absent but the score of the students will not be displayed. This can be illustrated using the following chart.  

Table 2: Showing the overall mark list as either present or absent 

 

 This system is implemented through AdaBoost classifier and Haar features. In this system, a GUI (graphical user interface) will be developed for storing the roll number and name of every student in the file. While the gathering of the student`s information is also creating dataset of the faces of the students and then storing them in a folder.  Immediately after the process of is completed trained images in the folder will be trained for the recognition of faces.  These are some crucial steps in this process of face recognition implementation.  For a real-time, scenario if the camera is put close to the doors of the classrooms it will continuously capture live images and this is done by capturing live streaming of the same camera.

 The images captured will be compared against the stored images in the dataset at the registration time. In case the images match the stored images then it will show the student`s roll number and name which has been recognized or stored. The information will be automatically stored in the attendance sheet together with the date and the name of the students.  And in case the image capture does not match the ones which are stored in the database then, it will be marked and stored under an unknown folder.

This system is very useful in recording the attendance of the students in colleges since it will store the attendance of the student together with the time and date.  Thus the members of faculties in colleges and universities will easily identify who attends the lecture on time. This suggested attendance system can also be employed in other fields like in laboratory attendance, organizational attendance, library attendance, and government office attendance. The block diagram below summarizes the working of the proposed attendance system;

 

Figure 8: Showing a block diagram for the working of the proposed attendance system (https://iopscience.iop.org/article/10.1088/1742-6596/2089/1/012078/pdf)  

 The block diagram in figure 8 above illustrates the flow of the working of the proposed system.  The first student will have to register all their information through the graphical user interface.   The information will be stored in the file after it develops the database of the student`s face and then trains it. AdaBoost classifier and Haar will be applied.  The system will capture a live image from the video stream and then store the file in the folder. All the matching images with a score of five and above out of ten trials will be marked present while those with a score less than five will be marked absent as illustrated in table 2 above.   

Different techniques for Image Processing

The system attendance dashboard for the student is illustrated in the following diagram

 

Figure 9: Showing the student dashboard for the proposed attendance system (https://iopscience.iop.org/article/10.1088/1742-6596/2089/1/012078/pdf) 

 From figure 9 above, the system dashboard shows a slot for entering the student`s name and ID, after feeding these details we must click the icon which is named take Image. After this the system will take the student`s image as illustrated in the following diagram;

 

Figure 10: Showing the system taking a photo (https://iopscience.iop.org/article/10.1088/1742-6596/2089/1/012078/pdf) 

 Through the use of the AdaBoost and the Haar feature, we need to classify and train the images.  There is also another icon which is named Trained image, thus after capturing the image, the student will have to click on the Train Image that will have to assign the name and the ID of the student directly as they have been fed to the system.  This can be illustrated in the following image of the captured image;

 

Figure 11: Showing the system where a student has been recognized (https://iopscience.iop.org/article/10.1088/1742-6596/2089/1/012078/pdf) 

 Another key step that needs to be undertaken is tracking images so that we click on the icon with the name Track Image.  Immediately after tracking a specific picture, it will indicate the student`s ID and name on the recognized image as illustrated in figure 12 below and after this, the information will get stored in the database.  After completion of the steps, we need to click on the Icon which is named Quit so that the student’s information which is stored as illustrated in the diagram below. And the system will generate an excel file that has been stored having the information of the student who is absent and also present in the class.

 

Figure 12: Showing the attendance of the student (https://iopscience.iop.org/article/10.1088/1742-6596/2089/1/012078/pdf)  

 The whole system can be understood through considering various scenarios which make it very simple;

1. When one student enters the classroom then the whole will system operate?

In case one student enters the room of the class at a time then the images of that specific student will be captured through a camera and then the image will have to be trained to recognize the student`s face on the basis of the image stored in the dataset.  On the recognition basis of the student`s face illustrated in table 2, the information is related to the specific image which will create an excel file and store the name of the student name, time, date, and roll number, the system will mark the student as either absent or present. After the data has been stored in this face recognition system, the aim of the system will give the attendance to the absent or present students.

2. A scenario of multiple classrooms of students entering simultaneously

 The system recognizes or detects multiple faces at a time several students enter the classroom as illustrated in the following diagram;

 

Figure 13: Showing the face of multiple recognized (https://iopscience.iop.org/article/10.1088/1742-6596/2089/1/012078/pdf) 

 After the images have been recognized, then the exit button is clicked and it will show the attendance notification in the box of attendance.   This attendance will be stored in the attendance list automatically together with the name, time, date, and ID of the students. 

Block diagram for the working and process of attendance system based on face recognition

Python programming language is also a good language which can help implemetatoon of this proposed attendance system.  A good python code which will be used for this can is given below;

# It helps in identifying the faces 

import cv2, sys, numpy, os 

size = 4 

haar_file = 'haarcascade_frontalface_default.xml' 

datasets = 'datasets' 

# Part 1: Create fisherRecognizer 

print('Recognizing Face Please Be in sufficient Lights...')

# Create a list of images and a list of corresponding names

(images, labels, names, id) = ([], [], {}, 0)

for (subdirs, dirs, files) in os.walk(datasets):

    for subdir in dirs:

        names[id] = subdir

        subjectpath = os.path.join(datasets, subdir)

        for filename in os.listdir(subjectpath):

            path = subjectpath + '/' + filename

            label = id

            images.append(cv2.imread(path, 0))

            labels.append(int(label)) 

        id += 1 

(width, height) = (130, 100) 

# Create a Numpy array from the two lists above 

(images, labels) = [numpy.array(lis) for lis in [images, labels]] 

# OpenCV trains a model from the images

# NOTE FOR OpenCV2: remove '.face' 

model = cv2.face.LBPHFaceRecognizer_create()

model.train(images, labels) 

# Part 2: Use fisherRecognizer on camera stream 

face_cascade = cv2.CascadeClassifier(haar_file) 

webcam = cv2.VideoCapture(0) 

while True: 

    (_, im) = webcam.read()

    gray = cv2.cvtColor(im, cv2.COLOR_BGR2GRAY)

    faces = face_cascade.detectMultiScale(gray, 1.3, 5)

    for (x, y, w, h) in faces: 

        cv2.rectangle(im, (x, y), (x + w, y + h), (255, 0, 0), 2)

        face = gray[y:y + h, x:x + w]

        face_resize = cv2.resize(face, (width, height)) 

        # Try to recognize the face 

        prediction = model.predict(face_resize) 

        cv2.rectangle(im, (x, y), (x + w, y + h), (0, 255, 0), 3)  

        if prediction[1]<500:  

           cv2.putText(im, '% s - %.0f' % 

(names[prediction[0]], prediction[1]), (x-10, y-10),

cv2.FONT_HERSHEY_PLAIN, 1, (0, 255, 0))

        else:

          cv2.putText(im, 'not recognized',

(x-10, y-10), cv2.FONT_HERSHEY_PLAIN, 1, (0, 255, 0)) 

    cv2.imshow('OpenCV', im)

    key = cv2.waitKey(10)

    if key == 27:

        break

# It helps in identifying the faces

import cv2, sys, numpy, os 

size = 4

haar_file = 'haarcascade_frontalface_default.xml'

datasets = 'datasets'

# Part 1: Create fisherRecognizer

print('Recognizing Face Please Be in sufficient Lights...')

# Create a list of images and a list of corresponding names

(images, labels, names, id) = ([], [], {}, 0) 

for (subdirs, dirs, files) in os.walk(datasets): 

    for subdir in dirs:

        names[id] = subdir

        subjectpath = os.path.join(datasets, subdir)

        for filename in os.listdir(subjectpath):

            path = subjectpath + '/' + filename

            label = id

            images.append(cv2.imread(path, 0))

            labels.append(int(label))

        id += 1

(width, height) = (130, 100)

# Create a Numpy array from the two lists above

(images, labels) = [numpy.array(lis) for lis in [images, labels]]

# OpenCV trains a model from the images

# NOTE FOR OpenCV2: remove '.face'

model = cv2.face.LBPHFaceRecognizer_create()

model.train(images, labels)

# Part 2: Use fisherRecognizer on camera stream

face_cascade = cv2.CascadeClassifier(haar_file)

webcam = cv2.VideoCapture(0)

while True:

    (_, im) = webcam.read()

    gray = cv2.cvtColor(im, cv2.COLOR_BGR2GRAY)

    faces = face_cascade.detectMultiScale(gray, 1.3, 5)

    for (x, y, w, h) in faces:

        cv2.rectangle(im, (x, y), (x + w, y + h), (255, 0, 0), 2)

        face = gray[y:y + h, x:x + w]

        face_resize = cv2.resize(face, (width, height))

        # Try to recognize the face

        prediction = model.predict(face_resize) 

        cv2.rectangle(im, (x, y), (x + w, y + h), (0, 255, 0), 3)

        if prediction[1]<500: 

           cv2.putText(im, '% s - %.0f' % 

(names[prediction[0]], prediction[1]), (x-10, y-10), 

cv2.FONT_HERSHEY_PLAIN, 1, (0, 255, 0)) 

        else:

          cv2.putText(im, 'not recognized', 

(x-10, y-10), cv2.FONT_HERSHEY_PLAIN, 1, (0, 255, 0))

    cv2.imshow('OpenCV', im)

    key = cv2.waitKey(10)

    if key == 27:

        break

The above code should be run after the model has been trained for the faces

When the codes are run before training the modes then the following errors are obtained;

 

For this project, the above python codes can be installed in a system like a laptop or a computer which will use Webcam to take the images of the students as they enter into their various classrooms. Through the use of the above python codes, the system will make it possible to take the faces and compare it with what is stored in the database of which if the face image is matching with the database it will be marked present but if not it will be marked as unknown.  This project will make it easier to mark attendance register since those students who are members of the class and their faces are not recognized during that time will be marked absent. Therefore, this proposed system makes marking easier and more reliable.

  The experimental arrangement is illustrated in the following diagram of the two databases. The face image assembling and the image which are mined geographic at the time procedure of the registration is conducted through the use of the database of the face recognition.   

 

Figure 14: Showing experimental setup (https://iopscience.iop.org/article/10.1088/1757-899X/263/4/042095/pdf) 

The software operation of the system is illustrated using the algorithm;

 

Figure 15: Showing the flowchart algorithm for the attendance system operation (https://iopscience.iop.org/article/10.1088/1757-899X/263/4/042095/pdf)  

 The algorithm which has been taken is given in the following processes for the algorithm;

1. Acquisition of the image
2. Normalization of the Histogram
3. Filtering of the noise
4. Skin classification

The first step in this process is to take an image through the use of the camera and there will be effects of lighting which will be in the captured image. And due to the dissimilar lighting condition and there will be some noise in the captured images as well then the image will have to be processed. The process includes the removal of the noise from the captured images, the elimination of this noise, the medium filter is done through area histogram normalization. The whole process is illustrated below;

  The acquisition of an image is also known as the image procurement; the image is settled which the pixel which is gathered. This can be illustrated using the following diagram;  

 

Figure 16: Showing appearance input  (https://iopscience.iop.org/article/10.1088/1757-899X/263/4/042095/pdf) 

 The image splitting which is taken through consumes brilliance else dimness which is employed as a good output of the images captured. The captured image will then get renovated grey picture the progress;

 

Figure 17: Showing class images in grey ( https://iopscience.iop.org/article/10.1088/1757-899X/263/4/042095/pdf) 

 

Figure 18: Showing  input for Histogram (https://iopscience.iop.org/article/10.1088/1757-899X/263/4/042095/pdf)  

 For the enhancement of dissimilarity in the histogram normalization of the spatial domain is a good technique to be used. This will assist in the identification of the learners on the rare. In this case a highly straightforwardly perceived. The obtained images will then get equalized as illustrated in the following diagram;  

 

Figure 19: Showing equalized images of the histogram (https://iopscience.iop.org/article/10.1088/1757-899X/263/4/042095/pdf) 

 There can be a lot of origin of the sounds in the participation of pictures when they were taken using a camera. There are lots of techniques that are available but one method enables classification of the frequent area which can be virtual but it may result in the removal of some key information in the captured images. An average cleaning of the image can be employed in this proposed attendance system where noise is dismissed through the splitting of normalized image.  This can be illustrated in the following diagram;

 

Figure 20: Showing noise filtration (https://iopscience.iop.org/article/10.1088/1757-899X/263/4/042095/pdf) 

 This technique is employed to promote the adeptness of the procedure of the identification of face, it represents a cutoff to effectiveness rise of the procedure of the identification of face.  The process of scanning of faces, jonnes procedure, and viola procedure are recycled for the exactness and discovering will be developed in case membrane is categorized.  The first Pixel will get systematically connected to transfigures and skin black and white and this is illustrated in the following diagram;

 

Figure 21: Showing classification of skin for the detection of faces   (https://iopscience.iop.org/article/10.1088/1757-899X/263/4/042095/pdf) 

 The technique of face detection realizes the segment faces through sphere shaping on the pictures which contain the learners` images. The techniques like the Haar classifiers are employed for face recognition. Immediately after conducting the skin classification, the degree of the algorithm will be enriched. The algorithm of the detection of the faces will be first employed for the changeability of the pictures having various actions and were captured for illumination conditions. This will be applied for the detection of the facial expression in a phase of audio-visual existent. The first procedure employed is for taking images and checking the functionality of capturing multiple images in a classroom. The next step will be the identification of the face images captured.  This procedure is recycled to rise the algorithm speed, The gathered images will then be allotted to bring a distinction drift. The identified images can be illustrated by the following images;

 

Figure 22: Showing face detected (https://iopscience.iop.org/article/10.1088/1757-899X/263/4/042095/pdf)  

Conclusion 

 In summary, the proposed system of attendance system using face recognition technology is for improving the attendance system in every field like in the organization, colleges, schools, and companies.  Capturing the live images cameras and applying various methods of face detection as well as the recognition of student faces in colleges will highly help in traditional and manual work.  In this proposed solution through the generation of the interface, we created the dataset.  We have to train the pictures through using Haar cascade and also through the use of AdaBoost classifier.  After training is completed the system will perfectly recognize and detect faces and non-faces. The images will be stored in the database will be compared against the stored images in the database. When the image gets recognized it will be getting updates together with the date and time.  Through storage of the images together with the date and time, it will be very easy for faculty in the colleges and universities to easily keep track of every student present in classes.  

References

Arsenovic, Marko, Srdjan Sladojevic, Andras Anderla, and Darko Stefanovic. "FaceTime—Deep learning based face recognition attendance system." In 2017 IEEE 15th International Symposium on Intelligent Systems and Informatics (SISY),IEEE, 2017.

Arulogun, O. T., Adeboye Olatunbosun, O. A. Fakolujo, and Olayemi Mikail Olaniyi. "RFID-based students attendance management system." International Journal of Scientific & Engineering Research 4, 2013.

Balcoh, Naveed Khan, M. Haroon Yousaf, Waqar Ahmad, and M. Iram Baig. "Algorithm for efficient attendance management: Face recognition based approach." International Journal of Computer Science Issues (IJCSI), 2012.

Barr, Jeremiah R., Kevin W. Bowyer, Patrick J. Flynn, and Soma Biswas. "Face recognition from video: A review." International journal of pattern recognition and artificial intelligence 26, 2012

BASAR, NURUL FARHANA BINTI. "EVENT ATTENDANCE SYSTEM USING ONE-TIME PASSWORD (OTP)." 2018.

Bolle, Ruud M., Jonathan H. Connell, Sharath Pankanti, Nalini K. Ratha, and Andrew W. Senior. Guide to biometrics. Springer Science & Business Media, 2013.

Chandramohan, J., R. Nagarajan, T. Dineshkumar, G. Kannan, and R. Prakash. "Attendance monitoring system of students based on biometric and gps tracking system." International Journal of Advanced engineering, Management and Science 3, 2017

Chintalapati, Shireesha, and M. V. Raghunadh. "Automated attendance management system based on face recognition algorithms." In 2013 IEEE International Conference on Computational Intelligence and Computing Research, IEEE, 2013.

Damanik, Rudolfo Rizki, Delima Sitanggang, Hendra Pasaribu, Hendrik Siagian, and Frisman Gulo. "An application of viola jones method for face recognition for absence process efficiency." In Journal of Physics: Conference Series, vol. 1007, no. 1, IOP Publishing, 2018.

Dhamecha, Tejas Indulal, Richa Singh, Mayank Vatsa, and Ajay Kumar. "Recognizing disguised faces: Human and machine evaluation." PloS one 9, 2014.

Dhole, Sampada A., and V. H. Patil. "Review of Multimodal Biometric Identification Using Hand Feature and Face." Bulletin of Electrical Engineering and Informatics 1, 2012.

Hutauruk, Sindak, Pandapotan Siagian, and Erick Fernando. "Development System With The Attendance of Content Based Image Retrieval (CBIR)." 2013

Jain, Anil K., and Ajay Kumar. "Biometric recognition: an overview." Second generation biometrics: The ethical, legal and social context, 2012

Kadry, Seifedine, and Mohamad Smaili. "Wireless attendance management system based on iris recognition." Scientific Research and essays 5, 2013.

Kakkad, Vishruti, Meshwa Patel, and Manan Shah. "Biometric authentication and image encryption for image security in cloud framework." Multiscale and Multidisciplinary Modeling, Experiments and Design 2, 2019.

Kar, Nirmalya, Mrinal Kanti Debbarma, Ashim Saha, and Dwijen Rudra Pal. "Study of implementing automated attendance system using face recognition technique." International Journal of computer and communication engineering, 2012.

Khan, Mubashir, and Md Zakariya. "Automated attendance system using face recognition." 2016.

Kumar, Ashu, Amandeep Kaur, and Munish Kumar. "Face detection techniques: a review." Artificial Intelligence Review 52, 2019

Li, Stan Z., and Anil Jain. Encyclopedia of biometrics. Springer Publishing Company, Incorporated, 2015.

Lukas, Samuel, Aditya Rama Mitra, Ririn Ikana Desanti, and Dion Krisnadi. "Student attendance system in classroom using face recognition technique." In 2016 International Conference on Information and Communication Technology Convergence (ICTC), IEEE, 2016.

Mahat, S. S., and S. D. Mundhe. "Proposed Framework: College attendance management system with mobile phone detector." international journal of research in IT and management 5, 2015

Manjani, Ishan, Snigdha Tariyal, Mayank Vatsa, Richa Singh, and Angshul Majumdar. "Detecting silicone mask-based presentation attack via deep dictionary learning." IEEE Transactions on Information Forensics and Security 12, 2017.

Masalha, Fadi, and Nael Hirzallah. "A students attendance system using QR code." International Journal of Advanced Computer Science and Applications 5, 2014.

Oloyede, Muhtahir O., and Gerhard P. Hancke. "Unimodal and multimodal biometric sensing systems: a review." IEEE access, 2016.

Padole, Chandrashekhar N., and Hugo Proenca. "Periocular recognition: Analysis of performance degradation factors." In 2012 5th IAPR international conference on biometrics (ICB), pp. 439-445. IEEE, 2012.

Patil, Ajinkya, and Mrudang Shukla. "Implementation of classroom attendance system based on face recognition in class." International Journal of Advances in Engineering & Technology 7, 2014

Savio, M. Maria Dominic, and S. Yuvaraj. "Attendance marking system using face recognition & rfid and prevention of examination malpractice system." International Journal of Engineering & Technology 7, 2018.

Senthilkumar, G., K. Gopalakrishnan, and V. Sathish Kumar. "Embedded image capturing system using raspberry pi system." International Journal of Emerging Trends & Technology in Computer Science 3, 2014.

Sim, Hiew Moi, Hishammuddin Asmuni, Rohayanti Hassan, and Razib M. Othman. "Multimodal biometrics: Weighted score level fusion based on non-ideal iris and face images." Expert Systems with Applications 41, 2014.

Siswanto, Adrian Rhesa Septian, Anto Satriyo Nugroho, and Maulahikmah Galinium. "Implementation of face recognition algorithm for biometrics based time attendance system." In 2014 International Conference on ICT For Smart Society (ICISS), 2014.

Srivastava, Tanya, Vanshika Vaish, Puneet Sharma, and Pooja Khanna. "Implementing Machine Learning for Face Recognition based Attendance Monitoring System." In 2019 6th International Conference on Computing for Sustainable Global Development (INDIACom), IEEE, 2019.

Sunaryono, Dwi, Joko Siswantoro, and Radityo Anggoro. "An android based course attendance system using face recognition." Journal of King Saud University-Computer and Information Sciences 33, 2021.

Vezzetti, Enrico, and Federica Marcolin. "3D human face description: landmarks measures and geometrical features." Image and Vision Computing 30, 2012.

Wagh, Priyanka, Roshani Thakare, Jagruti Chaudhari, and Shweta Patil. "Attendance system based on face recognition using eigen face and PCA algorithms." In 2015 International Conference on Green Computing and Internet of Things (ICGCIoT)IEEE, 2015.

Wang, Zhongyuan, Guangcheng Wang, Baojin Huang, Zhangyang Xiong, Qi Hong, Hao Wu, Peng Yi et al. "Masked face recognition dataset and application." arXiv preprint arXiv, 2020.

Zainal, Nur Izzati, Khairul Azami Sidek, Teddy Surya Gunawan, Hasmah Manser, and Mira Kartiwi. "Design and development of portable classroom attendance system based on Arduino and fingerprint biometric." In The 5th international conference on information and communication technology for the muslim world (ICT4M), IEEE, 2014.

Καπαντζ?κης, Ιω?ννης. "Developing a web based application for student attendance management." 2018