IMAGE PRE PROCESSING FOR TESSERACT OCR

SUGIHARTO, GOEI, STEVEN CHRISTIAN (2022) IMAGE PRE PROCESSING FOR TESSERACT OCR. Other thesis, Universitas Katholik Soegijapranata Semarang.

[img] Text
17.K1.0016-GOEI, STEVEN CHRISTIAN SUGIHARTO_COVER_a.pdf

Download (778kB)
[img] Text
17.K1.0016-GOEI, STEVEN CHRISTIAN SUGIHARTO_BAB I_a.pdf

Download (115kB)
[img] Text
17.K1.0016-GOEI, STEVEN CHRISTIAN SUGIHARTO_BAB II_a.pdf
Restricted to Registered users only

Download (126kB)
[img] Text
17.K1.0016-GOEI, STEVEN CHRISTIAN SUGIHARTO_BAB III_a.pdf

Download (115kB)
[img] Text
17.K1.0016-GOEI, STEVEN CHRISTIAN SUGIHARTO_BAB IV_a.pdf

Download (489kB)
[img] Text
17.K1.0016-GOEI, STEVEN CHRISTIAN SUGIHARTO_BAB V_a.pdf

Download (390kB)
[img] Text
17.K1.0016-GOEI, STEVEN CHRISTIAN SUGIHARTO_BAB VI_a.pdf

Download (113kB)
[img] Text
17.K1.0016-GOEI, STEVEN CHRISTIAN SUGIHARTO_DAPUS_a.pdf

Download (229kB)
[img] Text
17.K1.0016-GOEI, STEVEN CHRISTIAN SUGIHARTO_LAMP_a.pdf

Download (233kB)

Abstract

Sometimes processing text data or numbers in images, it makes us difficult to process the data. Ocr is software that converts text in image format or image files into text format that can be read and edited by computer applications, but sometimes there are also some that can't be detected And in my opinion through this pre processing will help the process of refinement or accuracy of this conversion process to a more accurate one, I use grayscale, then the image will go through the opening process where the image will be eroded first and then dilated, why don't I use the closing process, because what I want to detect here is text so that the results if using dilation will look worse than opening because it makes the writing close. I tried to use all pre-processing processes to find out which accuracy value was the best, where I compared the erosion, dilation, opening and closing processes. where the result is that dilation has the lowest value with 34% and the highest opening with 59% and that makes me use opening, I also compare that converters that go through pre-processing are higher than those that only use tesseract by comparison when using tesseract only get 43% while pre-processing is 59% more accurate

Item Type: Thesis (Other)
Subjects: 000 Computer Science, Information and General Works > 004 Data processing & computer science
Divisions: Faculty of Computer Science > Department of Informatics Engineering
Depositing User: mr AM. Pudja Adjie Sudoso
Date Deposited: 22 Mar 2022 04:54
Last Modified: 22 Mar 2022 04:54
URI: http://repository.unika.ac.id/id/eprint/28249

Actions (login required)

View Item View Item