CAHYONO, DANIEL SETIAWAN (2018) ALPHABETS IMAGE IDENTIFICATION USING ADVANCED LOCAL BINARY PATTERN AND CHAIN CODE ALGORITHM. Other thesis, Unika Soegijapranata Semarang.
|
Text (COVER)
14.K1.0003 Daniel Setiawan Cahyono (6.8%).COVER.pdf Download (1MB) | Preview |
|
|
Text (BAB I)
14.K1.0003 Daniel Setiawan Cahyono (6.8%).BAB I.pdf Download (72kB) | Preview |
|
![[img]](https://repository.unika.ac.id/style/images/fileicons/text.png) |
Text (BAB II)
14.K1.0003 Daniel Setiawan Cahyono (6.8%).BAB II.pdf Restricted to Registered users only Download (77kB) |
|
|
Text (BAB III)
14.K1.0003 Daniel Setiawan Cahyono (6.8%).BAB III.pdf Download (71kB) | Preview |
|
|
Text (BAB IV)
14.K1.0003 Daniel Setiawan Cahyono (6.8%).BAB IV.pdf Download (130kB) | Preview |
|
|
Text (BAB V)
14.K1.0003 Daniel Setiawan Cahyono (6.8%).BAB V.pdf Download (3MB) | Preview |
|
|
Text (BAB VI)
14.K1.0003 Daniel Setiawan Cahyono (6.8%).BAB VI.pdf Download (69kB) | Preview |
|
|
Text (DAFTAR PUSTAKA)
14.K1.0003 Daniel Setiawan Cahyono (6.8%).DAFTAR PUSTAKA.pdf Download (86kB) | Preview |
|
|
Text (LAMPIRAN)
14.K1.0003 Daniel Setiawan Cahyono (6.8%).LAMPIRAN.pdf Download (86kB) | Preview |
Abstract
ABSTRACT Optical Character Recognition (OCR) is the way computer process an image that contain some text and then try to find any font or number on that image, then convert it to digital text. Advanced Local Binary Pattern (ALBP) is a powerfull feature extraction that robust against illumination and rotation effect. The result of feature extraction ALBP is called histogram. Histogram of training image and testing image are compared using chi-square. The second algorithm is Chain Code. Chain Code is a feature extraction that used to detect the outline of object. The result of Chain Code is list of direction from the object. Before the feature extraction, the image needed to be normalized, the preprocessing consist of image resizing, convert image to grayscale, find the edge of the object and thinning the outline. The object of this research is the alphabet images. The training data contain of 26 image from a to z, the testing image are the alphabet images that already manipulated, for example rotated image, blurred image. The result of this research is the Chain Code algorithm had a better accuracy than the ALBP algorithm. And both of those algorithm could not identify the shape of alphabets if the image is blurred, or the image is noised. Keyword: ocr, albp, chain-code, alphabet, image identification.
| Item Type: | Thesis (Other) |
|---|---|
| Subjects: | 000 Computer Science, Information and General Works > 005 Computer programming, programs & data > Information Systems |
| Divisions: | Faculty of Computer Science > Department of Informatics Engineering |
| Depositing User: | Mr Lucius Oentoeng |
| Date Deposited: | 21 Jun 2018 04:21 |
| Last Modified: | 19 Jan 2021 06:42 |
| URI: | http://repository.unika.ac.id/id/eprint/16176 |
Actions (login required)
 |
View Item |
