Article Per Year (5 Year)
p-Index From 2020 - 2025
0.23
P-Index
This Author published in this journals
All Journal Jurnal Teknimedia: Teknologi Informasi dan Multimedia
Dinda Heidiyuan Agustalita
Unknown Affiliation
Computer Science & IT Control & Systems Engineering Engineering

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

 1 
IDENTIFIKASI PENULIS MENGGUNAKAN CONVOLUSIONAL NAURAL NETWORK BERDASARKAN KARAKTER TULISAN TANGAN Much Chafid; Muhammad Turmudzi; Agus Wibowo; Pratama Eskaluspita; Ervina Yuniati Rokhmah; Dinda Heidiyuan Agustalita
TEKNIMEDIA: Teknologi Informasi dan Multimedia Vol. 5 No. 2 (2024): Desember 2024
Publisher : Badan Penelitian dan Pengabdian Masyarakat (BP2M) STMIK Syaikh Zainuddin NW Anjani

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.46764/teknimedia.v5i2.209

Abstract

An application or system is needed to distinguish handwritten characters, and it can be used to identify author-based handwritten characters. The aim of this research is so that the system created can determine whether a piece of handwriting is the work of a particular author. The creation of handwritten text can be done by capturing the image using a scanner with an image quality of 300 dpi, segmenting it using the thresholding method and contour selection from the image, combining the segmented images and processing the image from the segmentation results. The results of the convolutional autoencoder can be input into transfer learning (lazy learning) using KNN method to match with the author's handwriting. The study used 100 data sets from 20 authors, each of whom wrote five times. In the first trial, we used the dataset in handwritten sentence fragments from the title of a poem by Chairil Anwar. Tests were performed by comparing the machine learning process with and without a convolutional autoencoder. The test results with a convolutional autoencoder showed an accuracy of 89%, while the results without a convolutional autoencoder showed an accuracy of 88%.
Co-Authors Agus Wibowo Ervina Yuniati Rokhmah Much Chafid Muhammad Turmudzi Pratama Eskaluspita