Face image quality inspection system according to ISO standard

Poor quality of face images is one of the reasons leading towards face recognition performance degradation. Therefore, this gives the motivation to have an International Standard (ISO/IEC 19794-5 Document) to provide guidelines for the usage of proper facial photographs for applications such as the...

Full description

Bibliographic Details
Main Author: Ong, Paik We (Author)
Format: Thesis
Published: 2012-06.
Subjects:
Online Access:Get fulltext
LEADER 01876 am a22001573u 4500
001 32076
042 |a dc 
100 1 0 |a Ong, Paik We  |e author 
245 0 0 |a Face image quality inspection system according to ISO standard 
260 |c 2012-06. 
520 |a Poor quality of face images is one of the reasons leading towards face recognition performance degradation. Therefore, this gives the motivation to have an International Standard (ISO/IEC 19794-5 Document) to provide guidelines for the usage of proper facial photographs for applications such as the E-Passport so that the face recognition process can be carried out more effectively. Nevertheless, despite the emergence of the ISO/IEC 19794-5 Standard, traditional passport photographs quality acceptance presently used is based on human visual perceptions and is very subjective. This is because there is no standardized checking system utilized which is able to give the same result on the quality acceptance of a photo regardless where or when it is being evaluated. As a result, there is a need for a system that is able to perform automatic checking on a passport size image to be developed. This project develops a system to perform automatic checking on a passport size image to ensure that it satisfies the image quality requirements according to the ISO/IEC 19794-5 Standard. The criteria considered in this project are image resolution, image aspect ratio, image brightness, image background colour, image eye distance, image head height and head width as well as image head rotation. The system developed managed to achieve at least 90% accuracy on all the attributes evaluated. 
546 |a en 
650 0 4 |a TK Electrical engineering. Electronics Nuclear engineering 
655 7 |a Thesis 
787 0 |n http://eprints.utm.my/id/eprint/32076/ 
856 |z Get fulltext  |u http://eprints.utm.my/id/eprint/32076/5/OngPaikwenMFKE2012.pdf