(For USM Staff/Student Only)

EngLib USM > Ω School of Electrical & Electronic Engineering >

Development of touch-less palm print biometrics for smart phone applications

Development of touch-less palm print biometrics for smart phone applications / Chan Yeow Pang
Biometrik merupakan sains dan teknologi yang mengukur dan menganalisis data biologi. Kebelakangan ini, telefon pintar telah menjadi semakin penting dalam kehidupan kita. Semakin banyak aplikasi telefon pintar yang bercadang untuk melaksanakan teknologi biometrik yang stabil dan praktikal. Sehubungan itu, biometrik tapak tangan dipilih untuk aplikasi telefon pintar disebabkan kelebihannya ke atas ciri-ciri biometrik lain. Bagaimanapun, terdapat beberapa cabaran dalam membangun sistem tanpa sentuhan seperti pemerolehan data tangan. Dalam projek ini, biometrik tanpa sentuhan yang berasaskan tapak tangan telah dibina untuk aplikasi telefon pintar dan algoritma pemerolehan data yang berasaskan penjejakan tangan, lembah puncak pengesanan bagi tangan dan pemilihan Kawasan Dikehendaki (ROI) telah diperkenalkan. Tiga model telefon pintar yang berbeza telah digunakan dalam projek ini. Imej tapak tangan telah diperoleh dengan menggunakan kamera telefon pintar dan kemudiannya dihantar ke server untuk pengesahan. Algoritma pemerolehan data telah digunakan untuk mengesan ciri-ciri tapak tangan. Seterusnya, Mesin Penyokong Vektor (SVM) dijadikan sebagai pengelas dalam pemadanan corak. Berikutannya, hasil pengesahan dihantar dari server kembali ke telefon pintar. Tiga sistem biometrik tapak tangan yang berasaskan bilangan data latihan yang berbeza telah dibina untuk penilaian prestasi. Lekungan Penerima Operasi (ROC) iaitu Kadar Penerimaan Keaslian (GAR) berbanding Kadar Penerimaan Kepalsuan (FAR) telah diplot dan Kadar Kesalahan Terseimbang (EER) telah dikira untuk menilai prestasi sistem-sistem tersebut. Projek ini telah membuktikan bahawa Samsung Galaxy Tab 2 dalam sistem biometrik yang berasaskan 20 data latihan memberikan prestasi terbaik dengan peratusan EER terendah. _______________________________________________________________________________________________________ Biometrics is the science and technology of measuring and analyzing biological data. In recent years, smart phones have imposed significant impacts to our life. More applications of smart phones aim to implement a reliable and practical biometric technology. Palm print biometrics appears promising for smart phone applications due its advantages over other types of biometric traits. However, there are some challenges in developing touch-less system such as the hand data acquisition. In this project, touch-less palm print biometrics is developed for smart phone applications and data acquisition algorithm which consists of hand tracking, peak valley detection and Region of Interest (ROI) selection is proposed. There are three different smart phone models involved. The palm print image is acquired by using the smart phone’s camera and then sent to server for verification. In the biometric system, data acquisition algorithm is applied to extract the palm print features. Following that, Support Vector Machine (SVM) is employed as classifier in pattern matching. Subsequently, the verification result is sent from server back to the smart phone. Three biometric systems with different number of training data are developed for performance evaluation. Receiver Operating Characteristic (ROC) curve of Genuine Acceptance Rate (GAR) versus False Acceptance Rate (FAR) is plotted and the Equal Error Rate (EER) is calculated to evaluate the system performances. In this project, it has been proven that Samsung Galaxy Tab 2 in the biometric system with 20 training data gives the best performance with the lowest EER percentage.
Contributor(s):
Chan Yeow Pang - Author
Primary Item Type:
Final Year Project
Identifiers:
Accession Number : 875005099
Language:
English
Subject Keywords:
Biometrics; technology; data
First presented to the public:
6/1/2013
Original Publication Date:
11/14/2019
Previously Published By:
Universiti Sains Malaysia
Place Of Publication:
School of Electrical & Electronic Engineering
Citation:
Extents:
Number of Pages - 103
License Grantor / Date Granted:
  / ( View License )
Date Deposited
2019-12-02 16:01:05.663
Submitter:
Mohd Jasnizam Mohd Salleh

All Versions

Thumbnail Name Version Created Date
Development of touch-less palm print biometrics for smart phone applications1 2019-12-02 16:01:05.663