Real time vision based attendance management system

Project Title

Real time vision based attendance management system

Project Area of Specialization Artificial IntelligenceProject Summary

The project is divided into two phases i) Recognition of individuals ii) Making a management system.

In first phase the system is trained to recognize the feature of each individual and store them in the database for comparison in later stages. The incoming data is matched with the already saved database for features matching and the best suited match is declared as the recognized image.

In the next phase the system will be capable of marking the attendance for which the individual should be sitting in the class 3 out of 4 times. The system will recognize the individual, calculate how many times that particular person was captures in the frame and on the basis of these results he/she is marked absent or present. Individuals who are not a part of the database will neither be marked present nor absent.

Project Objectives

An efficient and robust system to identify students for the purpose of marking their attendance in lecture room. First half of the system can detect and recognize the faces of the individuals in images. In later half the recognized images will be used for marking attendance of students in lecture room.Most of the work done on facial recognition is used for security purposes, now we need to extend this useful method in different fields too and this invention will be very helpful in it. In case of marking attendances in many institutes and colleges not only waste time of the students and teachers but also it has many shortcomings like proxy. So there is a need of a system that efficiently and accurately recognizes the faces automatically from the database without creating hassle and marks attendance. It will not only reduce the human effort but also save time and there wouldn’t be any confusion.

Project Implementation Method

Facial recognition technology is one of the most trending technologies marveling now days. It has become very important aspect of our life. It is an efficient and robust system. From a video or a picture we can detect if there’s a person or not. Then we can recognize the person who he/she is? So the basic three steps are:

i) Detection       ii) Feature extraction    iii) Recognition

The efficiency of our system mainly depends upon these three things. If the image is detected efficiently i.e. all the faces within an image are detected accurately then we have to extract the features. On the basis of the features extracted we recognize the person in the image. To improve the detection and recognition process there are different algorithms. Every algorithm has some pros and cons which makes it better but complex to compute. Feature extraction is very important task in the whole process. So to make an efficient system we have to choose the best algorithm.Once the features are extracted well we train our system. Repetitive training helps our system in better recognition of faces.

Faces are detected from the images using Viola Jones algorithm which has different steps that includes the extracting the facial feature using wavelets. in this process cascading is done which is very important as it reduces the computational work for the whole system. This algoritm detects the face on the basis that nose is slightly brighter than the rest of face and uper cheeks are brighter than the under eye.

after the correctdetection of face next step is to recognize the face which is done by using some pre-trained networks of convolutional neural network. CNN has different layers all of which have different functions to extract the features and train the model on given dataset. Accuracy of different networks are compared and the best is chosen. 

Benefits of the Project

As this system can be implemented in various places where there are plenty of people and we need to generate their attendance such as offices, classrooms, seminars so it can be installed in various places.

It not only reduces the human effort to a great extent but also saves time, removes any confusion and greatly reduces the chances for proxy to a major extent.

Technical Details of Final Deliverable

A GUI integrated with the camera which is installed in the lecture room and images are captured in the classoom and easily cassified using the gui provided to the user in the monitor

Final Deliverable of the Project HW/SW integrated systemType of Industry Education Technologies Artificial Intelligence(AI)Sustainable Development Goals Quality EducationRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	12000
Monitor	Miscellaneous	1	7000	7000
Raspberry pi	Equipment	1	5000	5000