PENGENALAN TULISAN STENOGRAFI MENGGUNAKAN CONVOLUTIONAL NEURAL NETWORK BERBASIS WEB
Total View This Week0
Total View This Week0
Institusion
Institut Teknologi Perusahaan Listrik Negara
Author
RAMADANI, ANDI PUTRI MAULANA
Ningrum, Rahma Farah
Agtriadi, Herman Bedi
Ningrum, Rahma Farah
Agtriadi, Herman Bedi
Subject
Teknik Informatika
Datestamp
2022-09-20 03:03:37
Abstract :
The workers, especially those who work as journalists and secretaries, have problems in recognizing shorthand writing. The need for media that can convert stenographic writing into a language that is easy to understand. Therefore, this study proposes to do stenographic image processing which then classifies the image as a supporting medium for journalists, secretaries, and other professionals in recognizing shorthand writing. The image is processed using the Convolutional Neural Network method, then the results are displayed in a website-based interface. The limit for the processed word category is 20 nouns. A total of 100 handwritings for each word class were collected so that the total images obtained were 2000 stenographic handwritten images. The results of image processing state that the average accuracy produced is 81%. From these results, it was concluded that the model formed using the Convolutional Neural Network method in shorthand writing was successfully recognized with a good level of accuracy.
The workers, especially those who work as journalists and secretaries, have problems in recognizing shorthand writing. The need for media that can convert stenographic writing into a language that is easy to understand. Therefore, this study proposes to do stenographic image processing which then classifies the image as a supporting medium for journalists, secretaries, and other professionals in recognizing shorthand writing. The image is processed using the Convolutional Neural Network method, then the results are displayed in a website-based interface. The limit for the processed word category is 20 nouns. A total of 100 handwritings for each word class were collected so that the total images obtained were 2000 stenographic handwritten images. The results of image processing state that the average accuracy produced is 81%. From these results, it was concluded that the model formed using the Convolutional Neural Network method in shorthand writing was successfully recognized with a good level of accuracy.
Download
book
BibTex
Latex, Jabref
cloud_download
Original Resource
url resource Institution
