Budi, Hadistya Wahyu Listyo (2026) SISTEM KLASIFIKASI PENYAKIT KANKER PARU-PARU BERBASIS CITRA CT-SCAN MENGGUNAKAN CONVOLUTIONAL NEURAL NETWORK (CNN) TERINTEGRASI CLINICAL DECISION SUPPORT. S1 thesis, Universitas PGRI Madiun.
![[img]](https://eprint.unipma.ac.id/style/images/fileicons/text.png) |
Text
HALAMAN DEPAN.pdf Download (1MB) |
|
Text
ABSTRAK.pdf Download (238kB) |
|
Text
BAB 1.pdf Download (287kB) |
|
Text
BAB 2.pdf Download (849kB) |
|
Text
BAB 3.pdf Restricted to Repository staff only Download (471kB) |
|
Text
BAB 4.pdf Restricted to Repository staff only Download (5MB) |
|
Text
BAB 5.pdf Download (272kB) |
|
Text
DAFTAR PUSTAKA.pdf Download (246kB) |
|
Text
LAMPIRAN.pdf Restricted to Repository staff only Download (7MB) |
Abstract
Budi. 2026. Lung Cancer Classification System Based on CT-Scan Images Using Convolutional Neural Network (CNN) Integrated with Clinical Decision Support. Thesis. Informatics Engineering Study Program, Faculty of Engineering, Universitas PGRI Madiun. Advisor (I) Saifulloh, S.Kom., M.Kom. Advisor (II) Puguh Jayadi, S.Kom., M.Kom. Lung cancer is a leading cause of global malignancy that requires early detection and appropriate treatment. The high visual complexity of CT-Scan images makes the scan results difficult for lay patients to understand directly without translation from medical experts, which can trigger psychological anxiety during the waiting period for medical consultation. This research develops a web-based intelligent self-screening system that integrates a Convolutional Neural Network (CNN) model with a Clinical Decision Support System (CDSS). The system adopts the ResNet50 architecture with a Two-Phase Transfer Learning strategy through a dual inference pipeline (dual-pipeline). The first pipeline is tasked with classifying the type of cancer (Adenocarcinoma, Large cell carcinoma, Squamous cell carcinoma, Normal) using the Chest CT-Scan Images dataset from Mendeley/Kaggle, while the second pipeline classifies the severity level (Malignant, Benign, Normal) using the IQ-OTH/NCCD dataset. The models were developed using the TensorFlow and Keras environments by applying Augmentation, oversampling, and class weighting techniques to address data imbalance. Model performance evaluation using a Confusion Matrix shows that the severity classification model achieved an accuracy of 88.69%, precision of 87.27%, recall of 88.69%, and an F1-score of 86.84%. Meanwhile, the cancer type classification model obtained an accuracy of 81.59%, precision of 84.63%, recall of 81.59%, and an F1-score of 81.78%. An Explainable AI (XAI) approach in the form of GradCAM was also implemented to transparently visualize the heatmap on lung anomaly areas. The system's frontend was built using React and Tailwind CSS, while the backend was developed using FastAPI to handle the model's real-time inference process. The black-box Testing results showed a functional success rate of 100%, proving that the system is stable and feasible to use for providing education, prescriptive medical action recommendations, and comprehensive health facility referral mapping.
| Item Type: | Thesis/Skripsi/Tugas Akhir (S1) |
|---|---|
| Kata Kunci: | Kanker Paru-Paru, Deep Learning, ResNet50, CT-Scan, Confusion Matrix, Clinical Decision Support System, React, FastAPI Lung Cancer, Deep Learning, ResNet50, CT-Scan, Confusion Matrix, Clinical Decision Support System, React, FastAPI |
| Subjects: | Q Science > Q Science (General) Q Science > QA Mathematics > QA75 Electronic computers. Computer science Q Science > QA Mathematics > QA76 Computer software |
| Divisions: | Fakultas Teknik > Teknik Informatika |
| Depositing User: | LISTYO WAHYU HADISTYA |
| Date Deposited: | 14 Oct 2026 04:54 |
| Last Modified: | 14 Oct 2026 04:54 |
| URI: | http://eprint.unipma.ac.id/id/eprint/10110 |
Actions (login required)
 |
View Item |
