Ardan, Indira Salsabila (2023) Rancang Bangun Sistem Deteksi Tumor Otak pada Citra MRI Menggunakan CNN. Other thesis, Institut Teknologi Sepuluh Nopember.
Text
05211940000046-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2025. Download (2MB) | Request a copy |
Abstract
Kanker otak termasuk kanker yang paling ganas dan, di Indonesia, terdapat 300 kasus tumor otak yang terdiagnosis setiap tahunnya. Untuk mendeteksi tumor otak, digunakan metode pencitraan seperti Magnetic Resonance Imaging (MRI). Namun, analisis manual citra MRI oleh dokter radiologi dapat menghasilkan hasil yang berbeda-beda, sehingga dapat terjadi inter-observer error dan hasil yang bergantung pada pengalaman dokter. Penelitian mengenai klasifikasi jenis tumor otak pada citra MRI juga masih terbatas. Oleh karena itu, dalam tugas akhir ini, akan dilakukan rancang bangun sistem untuk mendeteksi jenis-jenis tumor otak secara otomatis dan akurat pada citra MRI menggunakan metode Convolutional Neural Network (CNN) dan transfer learning. Tahapan tugas akhir ini meliputi studi literatur, pengumpulan data, pre-processing data, pengembangan arsitektur network, proses pelatihan model, uji coba & evaluasi model, implementasi model ke dalam aplikasi berbasis web, serta penulisan tugas akhir. Hasil dari tugas akhir ini adalah sebuah aplikasi berbasis web yang dapat mendeteksi jenis-jenis tumor otak pada citra MRI menggunakan framework Flask. Model yang digunakan menggunakan arsitektur CNN dengan base model ResNet50V2 yang sudah dilatih pada dataset ImageNet, head model yang terdiri dari satu lapisan fully connected dengan 512 nodes, dan lapisan output yang memprediksi input menjadi empat kelas citra MRI otak, yaitu normal, glioma, meningioma, dan pituitary. Untuk mencapai akurasi tertinggi, dilakukan pengaturan parameter yang tepat. Pada tugas akhir ini, digunakan algoritma optimasi Adam, dengan jumlah epoch 60 dan batch size 32. Selain itu, teknik ten-fold cross validation juga digunakan. Dengan menggunakan arsitektur yang diusulkan, diperoleh akurasi sebesar 95%.
=================================================================================================================================
Brain cancer is one of the most malignant cancers and, in Indonesia, there are 300 cases of brain tumors diagnosed each year. To detect brain tumors, imaging methods such as Magnetic Resonance Imaging (MRI) are used. However, manual analysis of MRI images by radiologists can produce different results, resulting in inter-observer error and results that depend on the doctor's experience. Research on the classification of brain tumor types on MRI images is also still limited. Therefore, in this final project, a system will be designed to detect brain tumor types automatically and accurately on MRI images using Convolutional Neural Network (CNN) and transfer learning methods. The stages of this final project include literature study, data collection, data pre-processing, network architecture development, model training process, model testing & evaluation, model implementation into web-based applications, and final project writing. The result of this final project is a web-based application that can detect types of brain tumors in MRI images using the Flask framework. The model uses CNN architecture with ResNet50V2 base model that has been trained on ImageNet dataset, head model consisting of one fully connected layer with 512 nodes, and output layer that predicts the input into four brain MRI image classes, namely normal, glioma, meningioma, and pituitary. To achieve the highest accuracy, appropriate parameter settings are made. In this paper, Adam's optimization algorithm is used, with the number of epochs 60 and batch size 32. In addition, ten-fold cross validation technique is also used. Using the proposed architecture, an accuracy of 95% was obtained.
Item Type: | Thesis (Other) |
---|---|
Uncontrolled Keywords: | Tumor otak, Magnetic Resonance Imaging (MRI), Convolutional Neural Network (CNN), Deep learning, Klasifikasi citra, Brain tumors, Image classification. |
Subjects: | T Technology > T Technology (General) T Technology > TA Engineering (General). Civil engineering (General) > TA1637 Image processing--Digital techniques |
Divisions: | Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information System > 57201-(S1) Undergraduate Thesis |
Depositing User: | Indira Salsabila Ardan |
Date Deposited: | 19 Jul 2023 04:47 |
Last Modified: | 19 Jul 2023 04:47 |
URI: | http://repository.its.ac.id/id/eprint/98608 |
Actions (login required)
View Item |