FACE RECOGNITION SYSTEM

We proposed a face recognition system for KKKUK, based on computer vision and proposed for solving security issues.  In our project, we applied significant experiments to test the performance of python using of Haar-Cascade Classifier for detecting face images. We analyzed this classifier

Project Title

FACE RECOGNITION SYSTEM

Project Area of Specialization

Artificial Intelligence

Project Summary

We proposed a face recognition system for KKKUK, based on computer vision and proposed for solving security issues.  In our project, we applied significant experiments to test the performance of python using of Haar-Cascade Classifier for detecting face images. We analyzed this classifier algorithm by dividing it into four steps. In this project, we have made the extraction of Haar-feature selection, create an integral image,   database training, and Cascading Classifiers to complete the detection step.  A Local Binary Patterns Histograms (LBPH) algorithm is used to complete the facial recognition system.  We proposed several parameters in LBPH and obtain a training dataset by applying a training algorithm. We completed the computational part by applying the LBPH operation and extracting the histograms.  Different kinds of law have been used to find the difference between two histograms like Euclidian distance, chi-square, etc. so this project gives 96% of accurate result than no one has done yet and it also detects and recognizes a face from the side also.

Project Objectives

The main concern is to detect and recognize the faces of pedestrians who are responsible for making bad use of the traffic control system in the university for finding the student, faculty, and unknown person.

We are attempting to construct a fast and efficient face recognition system that detects the faces of pedestrians very quickly in disordered backgrounds. We want to minimize the effects of the unwanted real-time environment by using the Haar-Cascade classifier. 

At one time the face detection is done; our following intent is to train our system with useful images. By using Local Binary Patterns for each image a feature vector is to be computed. With this feature vector, we desire to label the target images using Support Vector Machine (SVM) classification. 

Through our proposed method we want to reduce the rush caused by illegal people enter in university by adopting this proposal we can raise public awareness about minimizing the rush in this pandemic.

Project Implementation Method

For face recognition, at first, we have to register the user by detecting his/her face using the camera. By putting a unique user ID and image is separated from another image. 

The database contains a set of images and the person name of each image along with the user id. The database will be dynamic in the future.

Trainer create .yml file. This file contains all the matrices of each image. This matrix is used for comparing images.

The testing part is so excellent experience for us. In this part, we take 150 pictures of each person with many facial expressions for finding the weakness of our system.  But our system gives us very brilliant results. It can detect multiple faces in one frame and can successfully recognize multiple faces in one frame. It can – Detect faces b Recognize faces c Give Detail of a person who is recognized d Add a new user by putting a new user ID.    

Benefits of the Project

The main concern is to detect and recognize the faces of pedestrians who are responsible for making bad use of the traffic control system in the university for finding the student, faculty, and unknown person.

Facial recognition is quickly adopted around the world. Nowadays for various security purposes, facial recognition is used widely. Its popularity comes from the vast areas of potential applications. Some aspects that come from our projects are given below. 

The system can be applied at busy places like traffic signals, airports, and railway stations for detecting pedestrians' faces, when our system detects any suspicious faces it would make an internal alarm or signal. 

Technical Details of Final Deliverable

The testing part is so excellent experience for us. In this part, we take 150 pictures of each person with many facial expressions for finding the weakness of our system.  But our system gives us very brilliant results. It can detect multiple faces in one frame and can successfully recognize multiple faces in one frame. It can – Detect faces b Recognize faces c Give Detail of a person who is recognized d Add a new user by putting a new user ID.    

This test case is better than other related software which can detect and recognize faces. 

All the tested value was positive. We need to focus on the graphical interface of our system. 

The beta testing result was very supportive. End users will understand the system easily and acknowledgment will be positive.

Final Deliverable of the Project

HW/SW integrated system

Core Industry

Security

Other Industries

IT

Core Technology

Artificial Intelligence(AI)

Other Technologies

Others

Sustainable Development Goals

Sustainable Cities and Communities

Required Resources

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