(For USM Staff/Student Only)

EngLib USM > Ω School of Electrical & Electronic Engineering >

Development of intelligent doorbell featuring face recognition technology on raspberry pi board

Development of intelligent doorbell featuring face recognition technology on raspberry pi board / Goh Vi Nie
Pengesahan identiti adalah sangat penting dalam dunia zaman kini dan ia digunakan dalam kehidupan harian kita. Tugas untuk menjamin keselamatan maklumat atau harta fizikal menjadi semakin penting dan semakin susah. Dalam kebanyakan jenayah, penjenayah mengambil kesempatan atas kelemahan sistem kawalan biasa: sistem bukan mengizinkan akses dengan “siapa kita”, tetapi dengan “apa kita ada”, contohnya kad ID, kunci, kata laluan, atau nama ibu. Pengecaman muka adalah salah satu kaedah biometrik yang mempunyai dua kelebihan iaitu ketepatan yang tinggi dan gangguan yang rendah. Objektif projek ini adalah untuk mencipta sebuah loceng pintu pintar yang menggunakan teknologi pengecaman muka atas papan Raspberry Pi. Dalam proses pendaftaran, imej muka yang terdapat dalam set latihan ditukar kepada vektor muka. Selepas itu, purata vektor akan dikira sebelum dinormalisasikan untuk membuang ciri-ciri sama yang dikongsi oleh semua imej muka, supaya hanya ciri-ciri yang unik sahaja akan ditinggalkan. Selepas itu, matriks kovarians akan dikira untuk mendapatkan vektor eigen. Hasil keluaran latihan akan ditukar kepada fail xml. Proses pengesahan muka dilakukan dengan mengira nilai yakin peramalan di antara muka yang dikesan dengan imej muka di pangkalan data. Jika nilai tersebut lebih daripada ambang yang diletakkan, muka tersebut adalah disahkan. Satu perkakasan telah dibina untuk mempamerkan hasil keluaran sistem dan memperolehi maklumat input daripada pengguna bagi mengaktifkan sistem. Sistem ini berjaya mendapat keputusan pengesahan muka setinggi 81.25% dari 80 percubaan daripada 4 orang yang berbeza. Sebanyak 160 imej muka daripada 4 orang telah ditukar kepada imej potret dan disimpan dalam pangkalan data. _______________________________________________________________________________________________________ Identity recognition is very important in the world today and it is used in our everyday life. The need to maintain the security of information or physical property is becoming both increasingly important and difficult. In most of the crimes, the criminals were taking advantage of a fundamental flaw in the conventional access control systems: the systems do not grant access by "who we are", but by "what we have", such as ID cards, keys, passwords, PIN numbers, or mother's maiden name. Face recognition is one of the few biometric methods that possesses the merits of both high accuracy and low intrusiveness. The objective for this project is to create an intelligent doorbell incorporate with face recognition technology on a Raspberry Pi. In the enrolment process, the face images in the training set are converted to face vector. Then, the average face vector is calculated before normalization is carried out to remove the common features shared by all the face images, leaving only unique features for each image. Then, the covariance matrix is calculated to obtain the eigenvector. The training output will then be converted and compiled into an xml file when the process is completely finished. The identification process calculates the predicted confident value of the detected face with the faces saved in the database and if the value is more than an empirically defined threshold value, the face is identified. A user friendly hardware interface is built in order to show the system output result and also to obtain the input information from user to trigger the system. The identification success rate was 81.25% out of 80 attempts from 4 different people. There are a total of 160 face images from 4 people which are converted into portrait images to be saved in the database.
Contributor(s):
Goh Vi Nie - Author
Primary Item Type:
Final Year Project
Identifiers:
Accession Number : 875005801
Language:
English
Subject Keywords:
Identity; important; security
First presented to the public:
6/1/2015
Original Publication Date:
3/7/2019
Previously Published By:
Universiti Sains Malaysia
Place Of Publication:
School of Electrical & Electronic Engineering
Citation:
Extents:
Number of Pages - 87
License Grantor / Date Granted:
  / ( View License )
Date Deposited
2019-03-07 15:43:08.353
Submitter:
Mohd Jasnizam Mohd Salleh

All Versions

Thumbnail Name Version Created Date
Development of intelligent doorbell featuring face recognition technology on raspberry pi board1 2019-03-07 15:43:08.353