Smart Attendance and Monitoring System Using Face recognition

The system will automatically record student attendance in a classroom and offer staff, admin with the facilities so that they can maintain and track record. Face recognition is widely utilized in a variety of applications, including security cameras and CCTV systems, computer-human interaction, and

Project Title

Smart Attendance and Monitoring System Using Face recognition

Project Area of Specialization

Internet of Things

Project Summary

The system will automatically record student attendance in a classroom and offer staff, admin with the facilities so that they can maintain and track record. Face recognition is widely utilized in a variety of applications, including security cameras and CCTV systems, computer-human interaction, and access control systems used indoors, as well as network security. The problem of proxies and students being marked present even when they are not physically there can be easily fixed with this method.

It consists of 3 layers, front end, back-end, and database (cloud).

Even though these layers are dependent on each other, they have the ability to run on different places at different times. For example, the back-end that detects faces and recognizes them can be deployed anywhere and the front-end that displays data and gives the user the privilege to modify specific parts of it can be accessed by anyone from anywhere.

After the implementation of this system, we hope to minimize proxy chances that occur during lectures and our goal is to keep track of each student that enters or leaves the class. So not only this system will know when students are present but also when they're not, and the data that it'll generate will be easy for mining. The generated data will be available to find patterns among many different variables, for example, what's the relation between attending class regularly and the grades your get. During lectures proxies will be minimized using this system and an honest attendance report will be generated and sent to the teacher in class at the end of every lecture.

Project Objectives

Face recognition is one of the applications that will help saves time and effectively detects of eliminating risks of attendance by proxies. The algorithm is based on OpenCV Python libraries. This technology can be used in any sector where there is a need of monitoring or attendance. The objectives of the project are to keep track of and control attendance at universities and colleges. Objectives of this project are

• To reduce unnecessary time waste.

It will save the time required to interact with the system physically to mark the attendance.

• To automation security and attendance

There is no physical / user interaction required for the working of system.

• To reduce attendance proxies

As the system is fully automated it is impossible to hack or cheat the system.

• To strengthen the data integrity

As it is a cloud based system so the user data is secured with security algorithms of Mongodb.

 In addition to these, the core to these is to make attendance and monitoring totally automated. Once you run it there's no more need for human interaction. It does everything on its own from arranging lectures to marking students’ attendance and generating a report of everyone who was present and is present currently.

Project Implementation Method

Front-end and back-end are created using python as for database, mongodb is used. All 3 layers are linked and they can communicate with each other.

Front-end

Front-end is developed using PyQt5, PySide2 and tkinter. It uses a multi-threading interface. One thread is responsible for keeping the interface active at all times, other thread is responsible for refreshing the list of students along with other data that is coming from mongodb.

Database

Mongodb holds 3 type of tables

• Rooms tables hold the list of students currently present in each room.
• Report table holds the data required to generate a report for each faculty memeber.
• Schedule table holds schedule of each faculty member according to room numbers.

Back-end

Main script of back-end is checking current lecture continously. It makes a request to mongodb to see whose lecture is it right now. When a lecture starts it takes email of concerned faculy member and the room number that is supposed to be used for that lecture.

Then it starts the attendance process and at the end of the lecture, attendance report gets sent to concerned faculty member. 

Benefits of the Project

Main benefits of this project are

• Time-saving
• Proxies reduction
• Less or no human error
• Minimal user interaction
• Remote access to data

This project makes attendance monitoring easier than ever. Since no human interaction is required once it is placed on it's environment chances of human error are very low.

Technical Details of Final Deliverable

This system works on CNN and HOG model for face detection that are used by dlib (a python library). It uses dlib to perform general actions like face recognition, detection and matching. A multi-threading architecture is used to keep the system functional at all times meaning a dedicated thread is always handling a different aspect of this system.

 For example a thread is dedicated for handling the GUI and a different thread is dedicated for handling the refresh part. Refresh part is responsible for keeping the system up-to-date with changes made on database. An infinite loop continuously retrieves data from mongodb and displays it on the interface. Both of these actions occur at the same time in a different thread.

The way it handles face detection is by keeping a list of already existing data in memory and every time a person comes in front of camera it matches its face with the data it already has in memory. The data from local database gets loaded in memory at start.

The tolerance is kept 60% meaning it will bare a variation from original face for 60%. The reason behind using dlib instead of the usual haarcascade is that even though it needs higher hardware specs, it provides the best results and uses one of the best models for detection (CNN and HOG).

PyQt5 is used for designing the interface and that interface runs on a dedicated thread meaning it will stay responsive 100% of the time. In case an error occurs it will go in exception and the GUI will stay responsive throughout the process. It is created in python so the script that gets generated is able to be converted on any PC OS. There’s no restriction on keeping it windows only thing it will be able to run on any platform that supports personal computing environment.

Final Deliverable of the Project

HW/SW integrated system

Core Industry

Security

Other Industries

Education , IT , Others

Core Technology

Internet of Things (IoT)

Other Technologies

Artificial Intelligence(AI)

Sustainable Development Goals

Quality Education

Required Resources

If you need this project, please contact me on contact@adikhanofficial.com