Attire Checker and Face Recognition System

Chu, Caleb Ken Lun (2023) Attire Checker and Face Recognition System. Final Year Project (Bachelor), Tunku Abdul Rahman University of Management and Technology.

Text RSW_Caleb Chu Ken Lun_Fulltext.pdf Restricted to Registered users only Download (10MB)

Abstract

In this era of globalization, everyone emphasizes efficiency, quality, accuracy and speed. An inaccurate and inefficient system will be ousted and replaced by a better system. Currently, attire checking in TARUMT is still done manually by security guards which is ineffective and inefficient. The manual process requires a lot of manpower and cost to perform the tasks. Therefore, the main objective of this project is to develop an automated checking attire with facial recognition that can tackle the problem of manual checking attire system. The modules that make up the attire checker and face recognition system are object recognition, face recognition, clothing classification, blocked student list, and account management module. Other than increasing the efficiency of checking the attire of students' clothing in TARUMT, we can also raise students' awareness of their clothing. Students will take this issue seriously and adhere to the dress code because if they are denied access to campus due to improper attire, their class attendance will be affected. In this documentation, we will focus only on the object recognition, face recognition and account management modules. Object detection describes how our system recognizes people in a video, while face recognition describes how our system recognizes the identity of a person. The Account Management module describes the authority of management. For example, the administrator can create, update, delete and view the user data. Object recognition and detection is one of the useful and popular computer vision techniques that deals with locating and classifying objects. To make sure the system is error-free, our system will go through a number of tests, including unit testing, integration testing, and system testing. Therefore, this project will be able to solve the problem of manual checking, since most of the processes provided are automated, which can facilitate the work of security guards and minimize the manual workload.