GUTERRES, DAVID GINOLA (2026) CONVOLUTIONAL FEATURE OPTIMIZATION FOR LUNG DISEASE CLASSIFICATION ON IMAGE DATA. S1 thesis, Universitas Mercu Buana Jakarta.
|
Text (HAL COVER)
Cover.pdf Download (730kB) | Preview |
|
![[img]](https://repository.mercubuana.ac.id/style/images/fileicons/text.png) |
Text (BAB I)
Chapter I.pdf Restricted to Registered users only Download (108kB) |
|
|
Text (BAB II)
Chapter II.pdf Restricted to Registered users only Download (189kB) |
|
|
Text (BAB III)
Chapter III.pdf Restricted to Registered users only Download (188kB) |
|
|
Text (BAB IV)
Chapter IV.pdf Restricted to Registered users only Download (561kB) |
|
|
Text (BAB V)
Chapter V.pdf Restricted to Registered users only Download (29kB) |
|
|
Text (DAFTAR PUSTAKA)
References.pdf Restricted to Registered users only Download (96kB) |
|
|
Text (LAMPIRAN)
Attachment.pdf Restricted to Registered users only Download (317kB) |
Abstract
Lung diseases such as pneumonia, tuberculosis, and chronic obstructive pulmonary disease (COPD) present significant global health challenges, requiring accurate and efficient diagnostic methods. Convolutional Neural Networks (CNNs) have shown remarkable capabilities in medical image analysis, particularly in disease classification using chest X-rays and computed tomography (CT) scans. This study explores the optimization of CNN models for lung disease classification by implementing advanced preprocessing techniques, hyperparameter tuning, and dataset augmentation. Using publicly available datasets, various CNN architectures, including ResNet50, Efficient Net and Retina Net are evaluate to determine their effectiveness in distinguishing between different lung disease categories. The study achieves an optimized accuracy of up to 98.89%, demonstrating the potential of deep learning in assisting medical professionals with early and precise lung disease detection. Future research directions include integrating hybrid models, such as CNN-LSTM and ensemble learning techniques, to further improve classification accuracy and robustness. Keywords : (CNNs), X-rays, Computed Tomography (CT) scans, ResNet50, EfficientNet, RetinaNet, CNN-LSTM. Penyakit paru-paru seperti pneumonia, tuberkulosis, dan penyakit paru obstruktif kronis (PPOK) merupakan tantangan kesehatan global yang signifikan, yang memerlukan metode diagnostik yang akurat dan efisien. Jaringan Saraf Konvolusional (CNN) telah menunjukkan kemampuan luar biasa dalam analisis gambar medis, terutama dalam klasifikasi penyakit menggunakan sinar-X dada dan pemindaian tomografi komputer (CT). Studi ini mengeksplorasi optimasi model CNN untuk klasifikasi penyakit paru-paru dengan menerapkan teknik prapemrosesan canggih, penyesuaian hiperparameter, dan augmentasi dataset. Menggunakan dataset yang tersedia secara publik, berbagai arsitektur CNN, termasuk ResNet50, Efficient Net, dan Retina Net, dievaluasi untuk menentukan efektivitasnya dalam membedakan antara kategori penyakit paru-paru yang berbeda. Studi ini mencapai akurasi optimal hingga 98,89%, menunjukkan potensi pembelajaran mendalam dalam membantu tenaga medis dalam deteksi dini dan akurat penyakit paru-paru. Arah penelitian masa depan meliputi integrasi model hibrida, seperti CNN-LSTM dan teknik pembelajaran ensambel, untuk lebih meningkatkan akurasi dan ketahanan klasifikasi. Kata Kunci : (CNN), Sinar-X, Pemindaian Tomografi Terkomputasi (CT), ResNet50, EfficientNet, RetinaNet, CNN-LSTM.
Actions (login required)
 |
View Item |
