Nindiaputra, Rafi Anggara (2025) Optimasi Klasifikasi Penyakit Pneumonia Pada Citra X-Ray Menggunakan Metode KNN Dengan Ekstraksi Fitur Haralick Dan Local Binary Pattern. Undergraduate thesis, UPN Veteran Jawa Timur.
![[img]](https://repository.upnjatim.ac.id/style/images/fileicons/text.png) |
Text (Cover)
20081010242-cover.pdf Download (1MB) |
|
Text (bab 1)
20081010242-bab1.pdf Download (146kB) |
|
Text (bab 2)
20081010242-bab2.pdf Restricted to Repository staff only until 15 January 2028. Download (757kB) |
|
Text (bab 3)
20081010242-bab3.pdf Restricted to Repository staff only until 15 January 2028. Download (864kB) |
|
Text (bab 4)
20081010242-bab4.pdf Restricted to Repository staff only until 15 January 2028. Download (1MB) |
|
Text (bab 5)
20081010242-bab5.pdf Download (136kB) |
|
Text (daftar pustaka)
20081010242-daftarpustaka.pdf Download (143kB) |
|
Text (lampiran)
20081010242-lampiran.pdf Restricted to Repository staff only until 15 January 2028. Download (934kB) |
Abstract
The process of identifying pneumonia through lung X-ray images is still highly dependent on manual assessment by health workers, which takes a long time and is prone to human error. The limited number of radiologists and the high number of patients in health facilities make the pneumonia screening process inefficient. This study designed an automatic classification system to detect pneumonia in X-ray images by applying the KNN method combined with Haralick and Local Binary Pattern (LBP) feature extraction. The data used included 100 X-ray images from Husada Utama Hospital in Surabaya, covering normal and pneumonia categories, with a training and testing data split of [80:20 / 70:30 / 90:10]. Haralick feature extraction utilizes Gray Level Co-occurrence Matrix (GLCM) to obtain statistical texture features, while LBP is used to extract local texture features. This study optimized KNN parameters and compared classification performance using Haralick features alone, LBP alone, and a combination of both to achieve maximum accuracy. The results show that combining Haralick and LBP features provides higher classification accuracy than using each feature independently. The developed system achieved an accuracy of 80%, precision of 80%, recall of 80%, and an F1 score of 80% with an optimal k value of 5. This study confirms that optimizing the KNN method through a combination of feature extraction can improve effectiveness in detecting pneumonia. Keywords: Pneumonia, X-Ray Image, Haralick, Local Binary Pattern, KNN
| Item Type: | Thesis (Undergraduate) | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Contributors: |
|
||||||||||||
| Subjects: | T Technology > T Technology (General) T Technology > TK Electrical engineering. Electronics Nuclear engineering |
||||||||||||
| Divisions: | Faculty of Computer Science > Departemen of Informatics | ||||||||||||
| Depositing User: | Rafi A Rafi Nindiaputra | ||||||||||||
| Date Deposited: | 15 Jan 2026 06:54 | ||||||||||||
| Last Modified: | 15 Jan 2026 06:54 | ||||||||||||
| URI: | https://repository.upnjatim.ac.id/id/eprint/48734 |
Actions (login required)
 |
View Item |