SISTEM IDENTIFIKASI SIDIK JARI BERBASIS CONVOLUTIONAL NEURAL NETWORK

AULIA, AZMIN and Dwijayanti, Suci (2019) SISTEM IDENTIFIKASI SIDIK JARI BERBASIS CONVOLUTIONAL NEURAL NETWORK. Undergraduate thesis, Sriwijaya University.

[thumbnail of RAMA_20201_03041181520008.pdf] Text
RAMA_20201_03041181520008.pdf - Accepted Version
Restricted to Repository staff only
Available under License Creative Commons Public Domain Dedication.

Download (1MB) | Request a copy
[thumbnail of RAMA_20201_03041181520008_TURNITIN.pdf] Text
RAMA_20201_03041181520008_TURNITIN.pdf - Accepted Version
Restricted to Repository staff only
Available under License Creative Commons Public Domain Dedication.

Download (6MB) | Request a copy
[thumbnail of RAMA_20201_03041181520008_0030078404_01_front_ref.pdf]
Preview
Text
RAMA_20201_03041181520008_0030078404_01_front_ref.pdf - Accepted Version
Available under License Creative Commons Public Domain Dedication.

Download (419kB) | Preview
[thumbnail of RAMA_20201_03041181520008_0030078404_02.pdf] Text
RAMA_20201_03041181520008_0030078404_02.pdf - Accepted Version
Restricted to Repository staff only
Available under License Creative Commons Public Domain Dedication.

Download (317kB) | Request a copy
[thumbnail of RAMA_20201_03041181520008_0030078404_03.pdf] Text
RAMA_20201_03041181520008_0030078404_03.pdf - Accepted Version
Restricted to Repository staff only
Available under License Creative Commons Public Domain Dedication.

Download (121kB) | Request a copy
[thumbnail of RAMA_20201_03041181520008_0030078404_04.pdf] Text
RAMA_20201_03041181520008_0030078404_04.pdf - Accepted Version
Restricted to Repository staff only
Available under License Creative Commons Public Domain Dedication.

Download (184kB) | Request a copy
[thumbnail of RAMA_20201_03041181520008_0030078404_05.pdf] Text
RAMA_20201_03041181520008_0030078404_05.pdf - Accepted Version
Restricted to Repository staff only
Available under License Creative Commons Public Domain Dedication.

Download (72kB) | Request a copy
[thumbnail of RAMA_20201_03041181520008_0030078404_06_ref.pdf] Text
RAMA_20201_03041181520008_0030078404_06_ref.pdf - Bibliography
Restricted to Repository staff only
Available under License Creative Commons Public Domain Dedication.

Download (89kB) | Request a copy
[thumbnail of RAMA_20201_03041181520008_0030078404_07_lamp.pdf] Text
RAMA_20201_03041181520008_0030078404_07_lamp.pdf - Accepted Version
Restricted to Repository staff only
Available under License Creative Commons Public Domain Dedication.

Download (727kB) | Request a copy

Abstract

Fingerprints are unique identifiers which are not the same as others. Hence fingerprints can be used as biometrics, especially for security system. Various studies have been done to develop a biometric system using fingerprints. However, the proposed algorithms were not able to overcome the problems of high error rates and low accuracy segmentation. Therefore, this study focuses on implementing the Convolutional Neural Network (CNN) algorithm to identify fingerprints and improve the accuracy of identification. The research uses a primary data of 1484 which is taken from 106 registered students doing a practice in the Control and Robotics Laboratory, Department of Electrical Engineering, Faculty of Engineering, Sriwijaya University. Training dataset processed in the Convolutional Neural Network consists of the student's right and left thumbs, with the front-side capture position, right side, left side, upper and lower side of the thumb. In this study, the architecture designed of CNN uses convolutional layer, pooling layer, dropout and softmaxloss as classification, with a learning rate of 0.01, batch size 200, and dropout of 0.825. The testing is performed with new fingerprints which are different from the training data. There are three conditions used in the testing process, namely the initial condition, dropout condition and the condition of the batch normalization. The results showed that the system was able to identify fingerprints with an accuracy of 91.89% in initial conditions and 91.52% after added the dropout function. However, when the network is trained using batch normalization conditions, CNN's accuracy in identifying new data drops to 18.21%. This indicates that the batch normalization has not yet been able to generalize the new fingerprint data.

Item Type: Thesis (Undergraduate)
Uncontrolled Keywords: Convolutional Neural Network (CNN), Deep Learning, Image Recognition, Biometric, Fingerprint
Subjects: T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK1-9971 Electrical engineering. Electronics. Nuclear engineering > TK1 Electrical engineering--Periodicals. Automatic control--Periodicals. Computer science--Periodicals. Information technology--Periodicals. Automatic control. Computer science. Electrical engineering. Information technology.
Divisions: 03-Faculty of Engineering > 20201-Electrical Engineering (S1)
Depositing User: Users 2532 not found.
Date Deposited: 17 Oct 2019 07:40
Last Modified: 17 Oct 2019 08:16
URI: http://repository.unsri.ac.id/id/eprint/11892

Actions (login required)

View Item View Item