APLIKASI PENDETEKSI TUBERKULOSIS DAN PNEUMONIA MELALUI CITRA X-RAY MENGGUNAKAN MODEL MOBILENETV2 BERBASIS MOBILE SEBAGAI DECISION SUPPORT

Muhammad Alif Nadin Putra, . (2026) APLIKASI PENDETEKSI TUBERKULOSIS DAN PNEUMONIA MELALUI CITRA X-RAY MENGGUNAKAN MODEL MOBILENETV2 BERBASIS MOBILE SEBAGAI DECISION SUPPORT. Skripsi thesis, Universitas Pembangunan Nasional Veteran Jakarta.

[img] Text
ABSTRAK.pdf
Download (8MB)
[img] Text
AWAL.pdf
Download (4MB)
[img] Text
BAB 1.pdf
Restricted to Repository UPNVJ Only
Download (271kB)
[img] Text
BAB 2.pdf
Restricted to Repository UPNVJ Only
Download (845kB)
[img] Text
BAB 3.pdf
Restricted to Repository UPNVJ Only
Download (1MB)
[img] Text
BAB 4.pdf
Restricted to Repository UPNVJ Only
Download (1MB)
[img] Text
BAB 5.pdf
Download (253kB)
[img] Text
DAFTAR PUSTAKA.pdf
Download (254kB)
[img] Text
DAFTAR RIWAYAT HIDUP.pdf
Restricted to Repository staff only
Download (125kB)
[img] Text
LAMPIRAN.pdf
Restricted to Repository UPNVJ Only
Download (2MB)
[img] Text
HASIL PLAGIARISME.pdf
Restricted to Repository staff only
Download (30MB)
[img] Text
ARTIKEL KI.pdf
Restricted to Repository staff only
Download (416kB)

Abstract

Tuberculosis (TB) and Pneumonia are two pulmonary diseases with high morbidity and mortality rates, both globally and in Indonesia. Their detection generally still relies on the conventional interpretation of chest X-ray images, which heavily depends on radiologists, whose distribution in Indonesia remains uneven. This study aims to design and develop an Android mobile application named "Si Detek" as a decision-support tool for early screening of TB and Pneumonia through chest X-ray images, to measure the model's performance, and to evaluate the system's offline inference capability on mobile devices. The classification model was built using the MobileNetV2 Convolutional Neural Network architecture with three classes, namely Normal, Pneumonia, and TB. Training was conducted in two stages, feature extraction and fine-tuning, using the Adam optimizer and categorical crossentropy loss function. The trained model was converted into the float32 TensorFlow Lite format, reducing its size from 36 MB to 8.49 MB so that it runs entirely offline. The model achieved 94.08% accuracy, with macro-average precision, recall, and F1-score of 0.9450, 0.9408, and 0.9411, respectively. On a Xiaomi 25128RN17A, the average inference time was 105.24 ms with 215.25 MB memory usage. Black Box Testing passed 100%, and User Acceptance Testing by two radiographers rated the application as Good and Accepted. Therefore, "Si Detek" is proven capable of serving as a fast, practical, and efficient early screening tool for TB and Pneumonia.

Item Type: Thesis (Skripsi)
Additional Information: [No. Panggil: 2210511067] [Pembimbing: Neny Rosmawarni] [Pembimbing: Kraugusteeliana Kraugusteeliana] [Penguji 1: Ridwan Raafi'udin] [Penguji 2: Anis Fitri Nur Masruriyah]
Uncontrolled Keywords: Tuberculosis, Pneumonia, X-ray image, MobileNetV2, mobile.
Subjects: Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Divisions: Fakultas Ilmu Komputer > Program Studi Informatika (S1)
Depositing User: MUHAMMAD ALIF NADIN PUTRA
Date Deposited: 27 Aug 2026 07:58
Last Modified: 27 Aug 2026 07:59
URI: http://repository.upnvj.ac.id/id/eprint/51662

Actions (login required)

View Item View Item