Face Recognition Based Automated Attendance Portal

Project Title

Face Recognition Based Automated Attendance Portal

Project Area of Specialization Artificial IntelligenceProject Summary

Over the past decade, taking down students' attendance process has been developed and changed. The driving force of this development is the desire to automate, facilitate, speed up and save time and efforts. Although the attendance systems are around us everywhere, our university lecturers still use a traditional way to record students’ attendance by calling out students’ names. Which is time consuming and associated with high error scales in this project we use facial recognition technology to record students attendance by using a camera device connected to the internet. A person's face has distinctive bodily form and traits which are used to perceive or confirm a person. Facial recognition records the information of the face; special face recognition techniques measure the biometric of the face. Different face recognition methods measure the biometric of the face. Facial recognition has become a very important topic in recent years. Facial recognition is effectively applied in various applications like security systems, authentication, entrance control, surveillance system, unlocking of smartphones and social networking systems, etc. Most of the practices do not use facial recognition as the main form of conceding entry. However, with advancement in tech and science facial recognition technology has the ability to change the general way of character identification.

Project Objectives

The objective of this project is to develop a face recognition based automated attendance system.Most lecturers have a significant number of students and it is hard to keep taking or tracking all their absence. Facial recognition is commonly used in many institutions to take attendance of a significant number of students. There are many errors that could occur during this process, including misidentification and self-recognition. Lecturer can control the errors and correct them.

Expected goals in order to achieve the objectives are:

1. Our primary goal is to help the lecturers, improve and organize the process of track and manage student attendance and absenteeism using face recognition technology.

2. To detect the face segment from the video frame.

3. To extract the useful features from the face detected.

4. To classify the features in order to recognize the face detected.

5. To record the attendance of the identified student.

Figure 1.2 Block Diagram of the General Framework

Project Implementation Method

The approach performs a face recognition based student attendance portal. The approach flow diagram begins with the capture of video by camera device using Media Capture and Streams API. The video stream frames will be feeded to the face-api.js which will extract  features from the facial capture frame with subjective selection. Face-api.js CNN, which returns the 68 point face landmarks of a given face image and creates a face description of the facial images to be recognized. The recognized face description will be matched with the detected facial from the video stream and its attendance will be marked in mysql database using PHP. The attendance report will be shown to users in web GUI by HTML, Javascript, bootstrap and ajax.

The flow chart for the proposed system is shown in Figure 3.1 respectively. 

 Figure 3.1 Flow chart Diagram of Automated Attendance Portal approach

Benefits of the Project

The project titled “Face Recognition Based Automated Attendance Portal” covers the scope of the present system. It is a web based system with the capability to take attendance  from video streams, track unknown bodies and easily check the attendance of students. The main scope of this project is to verify the regulation of students in attending lectures and identify unknown buddies. It focuses only on checking the attendance of students. This study covers purely the formulation of the proposed system as a replacement for the present system. The development of this system is going to be very important to every Department because of applying attendance easily and systematically. Using this system will reduce the absenteeism mistakes of the students and recognize unknown people. This system is for the benefit of all Departments because it is used to determine if the student attends or not.

Technical Details of Final Deliverable

We are going to develop an automated  attendance system that will mark attendance and verify unknown students by face recognition.The face will be detected using SSD (Single Shot Multibox Detector), which is basically a CNN based on MobileNet V1, with some additional box prediction layers stacked on top of the network. The networks return the bounding boxes of each face, with their corresponding scores. The bounded box will be aligned to the center of the face before passing them to the face recognition network, as this will make face recognition much more accurate. MTCNN (Multi-task Cascaded Convolutional Neural Network) implementation, which is mostly around nowadays for experimental purposes however.The networks return the bounding boxes of each face, with their corresponding scoresI want to point out that we want to align the bounding boxes, such that we can extract the images centered at the face for each box before passing them to the face recognition network, as this will make face recognition much more accurate. For that purpose face-api.js implements a simple CNN, which returns the 68 point face landmarks of a given face image.

From the landmark positions, the bounding box can be centered on the face. In the following you can see the result of face detection (left) compared to the aligned face image (right)

Now we can feed the extracted and aligned face images into the face recognition network, which is based on a ResNet-34 like architecture and basically corresponds to the architecture implemented in dlib. The network has been trained to learn to map the characteristics of a human face to a face descriptor (a feature vector with 128 values), which is also oftentimes referred to as face embeddings.

We will use the face descriptor of each extracted face image and compare them with the face descriptors of the reference data. More precisely, we can compute the euclidean distance between two face descriptors and judge whether two faces are similar based on a threshold value (for 150 x 150 sized face images 0.6 is a good threshold value). Using euclidean distance works surprisingly well, but of course you can use any kind of classifier of your choice. The following gif visualizes the comparison of two face images by euclidean distance. Which will verify known and unknown students with labels to mark their attendance and recognize them.

Final Deliverable of the Project HW/SW integrated systemCore Industry EducationOther Industries Security Core Technology Artificial Intelligence(AI)Other TechnologiesSustainable Development Goals Quality EducationRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	78500
Camera	Equipment	1	24500	24500
LCD Screen	Equipment	1	17000	17000
Wi-Fi Device	Equipment	1	16000	16000
Storage Devices	Equipment	1	11000	11000
Other	Miscellaneous	10000	1	10000