Skripsi

PERBAIKAN KUALITAS CITRA DENGAN NONLINEAR ACTIVATION FREE NETWORK UNTUK KLASIFIKASI PENYAKIT ALZHEIMER PADA CITRA MRI

---

Alzheimer’s disease is a progressive neurodegenerative disorder that is the leading cause of dementia. The use of Magnetic Resonance Imaging (MRI) in ResNet18-based classification can be hindered by the presence of noise, particularly Rician noise, which obscures important structural details and reduces model performance. This study employs the Nonlinear Activation Free Network (NAFNet) to improve MRI image quality and evaluate classification performance across three data conditions, namely raw images, noisy images (generated by synthetic noise), and enhanced images. Image quality is evaluated using PSNR and SSIM, while classification performance is measured using accuracy and macro F1-score to account for class imbalance. The results show that noise significantly reduces performance from an accuracy of 98.34% and macro F1-score of 98.87% on raw images to 79.11% and 76.82% on noisy images. After enhancement using NAFNet, image quality improved, with PSNR increasing from 25.91 dB to 34.57 dB and SSIM from 0.5361 to 0.9626, while classification performance improved to an accuracy of 96.78% and macro F1-score of 97.61%. These results demonstrate that image enhancement improves the model’s robustness against noise degradation in MRI images.

---

Availability

Detail Information

Series Title

-

Call Number

T1962072026

Publisher

Indralaya : Prodi Teknik Informatika, Fakultas Ilmu Komputer Universitas Sriwijaya., 2026

Collation

xviii, 56 hlm.; ilus.; 29 cm

Language

Indonesia

ISBN/ISSN

-

Classification

006.370 7

Content Type

Text

Media Type

-

Carrier Type

-

Edition

-

Subject(s)

Pengolahan Citra Digital
Prodi Teknik Informatika

Specific Detail Info

-

Statement of Responsibility

TUTI

Other version/related

File Attachment

Comments

---

You must be logged in to post a comment