PENGENALAN POLA TANDA TANGAN MENGGUNAKAN HISTOGRAM OF ORIENTED GRADIENTS DAN BACKPROPAGATION

Budiyanto, Rendi (2020) PENGENALAN POLA TANDA TANGAN MENGGUNAKAN HISTOGRAM OF ORIENTED GRADIENTS DAN BACKPROPAGATION. Tugas Akhir thesis, University of Technology Yogyakarta.

[img]
Preview
Text
Naskah Publikasi-5130411332 Rendi Budiyanto.pdf

Download (809kB) | Preview

Abstract

Signature is a means of identifying someone. The existence of a signature in a document indicates that the party who signed the document already knows and agrees to the contents of a document. Signature identification is important in the validation of a document. Visually recognizing a signature is quite easy to do. However, if there are a large number of documents that are validated based on the signatures of the signed parties, this will take more time. Backpropagation is an algorithm that can be used to recognize, classify, and predict data. The backpropagation algorithm learning process is included in the category of supervised learning methods. In this study, backpropagation is combined with batch normalization to improve its performance. The dataset used was obtained from CEDAR from 55 respondents with 24 original data and 24 fake data, the second dataset was photos of 5 respondents each 16 original data and 16 fake data. Preprocessing is done to adjust the size to 100x400 pixels. Feature extraction using the Histogram of Oriented Gradients with bin 9, cells 20x20, and 1x1 block which yields 900 vectors. Testing is done by taking 25% of each dataset, 660 data from the first dataset and 40 for the second dataset. The results obtained from several experiments, for the first dataset, the highest was 91.36 %, while the second dataset was 87.5% Keywords: Signature, Histogram of Oriented Gradients, Backpropagation, Batch Normalization

Item Type: Thesis (Skripsi, Tugas Akhir or Kerja Praktek) (Tugas Akhir)
Subjects: T Technology > T Technology (General)
Divisions: Fakultas Sains Dan Teknologi > S1 Informatika
Depositing User: Kaprodi S1 Informatika UTY
Date Deposited: 12 Dec 2020 05:06
Last Modified: 12 Dec 2020 05:06
URI: http://eprints.uty.ac.id/id/eprint/6245

Actions (login required)

View Item View Item