IDENTIFIKASI IRIS MATA MENGGUNAKAN BIOMIMETIC PATTERN RECOGNITION
Total View This Week0
Total View This Week0
Institusion
Universitas Sriwijaya
Author
DIO GHIFFARI (STUDENT ID : 09121002026)
Samsuryadi Samsuryadi (LECTURER ID : 0004027101)
Anggina Primanita (LECTURER ID : 0206088901)
Samsuryadi Samsuryadi (LECTURER ID : 0004027101)
Anggina Primanita (LECTURER ID : 0206088901)
Subject
T58.5-58.64 Information technology
Datestamp
2019-09-11 01:47:30
Abstract :
The recognition of the iris image pattern is to identify based on the unique patterns of iris in an iris image. The iris image identification process is performed using a number of steps such as edge detection with Prewitt operator, detecting iris circle by Cyrcle Hough Transform method, normalizing iris image with Daughman's Rubber sheet model, and feature extraction using Independent Component Analysis. The result of feature extraction is used for identification process by Biomimetic Pattern Recognition (BPR) method. Based on the topology used, BPR builds the scopes in high-dimensional space using theoretical geometry of knowledge and image thinking resulting in scoping on the nerve network with many weights. The research was conducted by 2 stages, identifying 80 individual samples, each of which had 5 iris images with 4 training images and 1 test image having average success of 71,67%. While the identification of 25 samples for 3 image training and 2 test images with an average of 86% success.
The recognition of the iris image pattern is to identify based on the unique patterns of iris in an iris image. The iris image identification process is performed using a number of steps such as edge detection with Prewitt operator, detecting iris circle by Cyrcle Hough Transform method, normalizing iris image with Daughman's Rubber sheet model, and feature extraction using Independent Component Analysis. The result of feature extraction is used for identification process by Biomimetic Pattern Recognition (BPR) method. Based on the topology used, BPR builds the scopes in high-dimensional space using theoretical geometry of knowledge and image thinking resulting in scoping on the nerve network with many weights. The research was conducted by 2 stages, identifying 80 individual samples, each of which had 5 iris images with 4 training images and 1 test image having average success of 71,67%. While the identification of 25 samples for 3 image training and 2 test images with an average of 86% success.
File :
RAMA_55201_09121002026.pdf
RAMA_55201_09121002026_0004027101_0206088901_02.pdf
RAMA_55201_09121002026_0004027101_0206088901_03.pdf
RAMA_55201_09121002026_0004027101_0206088901_04.pdf
RAMA_55201_09121002026_0004027101_0206088901_05.pdf
RAMA_55201_09121002026_0004027101_0206088901_06_ref.pdf
RAMA_55201_09121002026_0004027101_0206088901_07_lamp.pdf
RAMA_55201_09121002026_0004027101_0206088901_01_front_ref.pdf
Download
book
BibTex
Latex, Jabref
cloud_download
Original Resource
url resource Institution
