Pengenalan Pola Pembuluh Darah (Vasculature) Pada Sklera Menggunakan Image Processing Berbasis Deep Learning.

Malau, Josafat Mangasitua (2022) Pengenalan Pola Pembuluh Darah (Vasculature) Pada Sklera Menggunakan Image Processing Berbasis Deep Learning. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 02311540000068-Undergraduate_Thesis.pdf Restricted to Repository staff only Download (3MB)

Abstract

Pengenalan pola pembuluh darah pada sklera merupakan salah satu alternatif penciri biometrik okular yang unik dan berbeda untuk setiap individu selain iris yang dapat digunakan untuk melakukan kegiatan autentikasi atau pengenalan identitas. Convolutional Neural Network (CNN) telah terbukti memiliki performa yang lebih baik daripada metode konvensional dan menghasilkan akurasi yang lebih tinggi dalam melakukan kegiatan segmentasi maupun pengenalan citra. Pengenalan pola pembuluh darah pada sklera dilakukan dengan dua tahap segmentasi yaitu segmentasi sklera dan pembuluh darah (vasculature) pada sklera menggunakan deep learning dengan arsitektur SegNet dan arsitektur Deep Neural Network (DNN) untuk pengenalan identitas menggunakan pola pembuluh darah pada sklera. Pada penelitian ini digunakan dataset Sclera Blood Vessels, Periocular and Iris (SBVPI) sebagai data training, validasi, dan uji coba. Segmentasi sklera yang dilakukan menghasilkan nilai akurasi 98.7%, precision 93.0%, dan recall 92.1%. Segmentasi vasculature pada sklera menghasilkan nilai akurasi 75.7%, precision 78.8%, dan recall 75.3%. Pengenalan pola vasculature menghasilkan nilai akurasi 94.2%, precision 98.3%, dan recall 90.2%.
===================================================================================================================================
Sclera vasculature pattern recognition is one of the alternative ocular biometric markers that are unique and different for everyone apart from iris which can be used to perform authentication or identity recognition activities. Convolutional Neural Network (CNN) has been proven to have better performance than conventional methods and acquire higher accuracy in segmentation and image recognition activities. The sclera vasculature pattern recognition is carried out in two stages of segmentation, namely segmentation of sclera region and vasculature region using deep learning with SegNet and Deep Neural Network (DNN) for identity recognition using the pattern of sclera vasculature. In this study, Sclera, Blood Vessels, Periocular, and Iris (SBVPI) dataset was used as training, validation, and testing processes. The scleral segmentation performed acquires 98.7% accuracy, 93.0% precision, and 92.1% recall. Scleral vasculature segmentation acquires 75.7% accuracy, 78.8% precision, and 75.3% recall. Sclera vasculature recognition acquires 94.2% accuracy, 98.3% precision, and 90.2% recall.

Item Type:	Thesis (Other)
Additional Information:	RSF 006.3 Mal p-1 2022
Uncontrolled Keywords:	Biometrik Okular, Sklera, Vasculature, CNN, DNN, SegNet, Deep learning.Ocular Biometric, Sclera, Vasculature, CNN, DNN, SegNet, Deeplearning.
Subjects:	T Technology > TE Highway engineering. Roads and pavements > TE228.37 Vehicular ad hoc networks (Computer networks)
Divisions:	Faculty of Industrial Technology and Systems Engineering (INDSYS) > Physics Engineering > 30201-(S1) Undergraduate Thesis
Depositing User:	Mr. Marsudiyana -
Date Deposited:	11 May 2026 01:21
Last Modified:	11 May 2026 01:21
URI:	http://repository.its.ac.id/id/eprint/133095

Actions (login required)

View Item