Firdawati, Nina Sajidah and Dewi Anggraeni, S.Si., M.Si (2024) Pengaruh Region of Interest (ROI) terhadap Akurasi Deteksi Cacat Color Tone Botol Susu dengan Metode Ekstraksi Ciri Gray Level Co-Occurrence Matrix (GLCM) dan Model Decision Tree. Sarjana thesis, Universitas Brawijaya.
Abstract
Luasan Region of Interest (ROI) dapat mempengaruhi statistik gambar dan akurasi model decision tree. Semakin kecil luasan ROI yang digunakan, maka noise pada citra akan semakin berkurang, mempengaruhi nilai pengukuran beberapa parameter Gray Level CoOccurrence Matrix (GLCM) dan keterpisahan data dalam analisis GLCM, yang berdampak pada klasifikasi decision tree. Dalam penelitian ini, penggunaan "printing only ROI" menghasilkan nilai tertinggi pada beberapa parameter GLCM dibandingkan dengan "full image ROI" dan "bottle ROI", serta menunjukkan keterpisahan data yang lebih baik dalam beberapa parameter GLCM. Parameter tersebut meliputi dissimilarity, energy, homogeneity, dan contrast. Analisis GLCM ini dilakukan setelah proses thresholding untuk mengurangi noise yang kemudian hasilnya digunakan untuk membangun model decision tree. Akurasi yang diperoleh dari penelitian ini adalah 92.98% untuk full image ROI, 94.04% untuk bottle ROI, dan 98.25% untuk printing only ROI.
English Abstract
Region of Interest (ROI) size can affect image statistics and the accuracy of decision tree models. Smaller ROI size, can minimize the noise in the image, affecting the measurement values of several parameters in the Gray Level Co-Occurrence Matrix (GLCM) and the data separation in GLCM analysis, which impacts decision tree classification. In this study, the use of "printing only ROI" yielded the highest values for several GLCM parameters compared to "full image ROI" and "bottle ROI," indicating better data separation. These parameters include dissimilarity, energy, homogeneity, and contrast. GLCM analysis was performed after thresholding to reduce noise and was used to build the decision tree model. The obtained accuracies from this study were 92.98% for full image ROI, 94.04% for bottle ROI, and 98.25% for printing only ROI.
Item Type: | Thesis (Sarjana) |
---|---|
Identification Number: | 052409 |
Uncontrolled Keywords: | Decision tree, Gray Level Co-Occurrence Matrix (GLCM), Region of interest (ROI) |
Divisions: | Fakultas Matematika dan Ilmu Pengetahuan Alam > Instrumentasi Fisika |
Depositing User: | Unnamed user with username nova |
Date Deposited: | 22 Aug 2024 04:17 |
Last Modified: | 22 Aug 2024 04:17 |
URI: | http://repository.ub.ac.id/id/eprint/223690 |
![]() |
Text (DALAM MASA EMBARGO)
Nina Sajidah Firdawati.pdf Restricted to Registered users only Download (5MB) |
Actions (login required)
![]() |
View Item |