Aplikasi Berbasis Web Untuk Mendeteksi Parasit Plasmodium Pada Gambar Apusan Darah Dengan Pendekatan Deep Learning

Yuliawan, Tri Rizky (2023) Aplikasi Berbasis Web Untuk Mendeteksi Parasit Plasmodium Pada Gambar Apusan Darah Dengan Pendekatan Deep Learning. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 05311940000024-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 September 2025. Download (4MB) | Request a copy

Abstract

Malaria adalah masalah kesehatan global yang serius yang menyebabkan morbiditas dan mortalitas yang signifikan. Diagnosis malaria yang akurat dan tepat waktu sangat penting untuk pengobatan dan pengendalian penyakit yang efektif. Dalam penelitian ini, penulis mengeksplorasi beberapa pendekatan deep learning untuk mendeteksi parasit plasmodium parasit penyebab penyakit malaria. Dataset yang akan digunakan adalah Thin Blood Smears Dataset yang disediakan oleh U.S National Library of Medicine. Penulis membandingkan dan mengevaluasi kinerja teknik transfer learning dan snapshot ensemble pada dataset yang kemudian akan dibantu dengan Gradient Camera. Transfer learning digunakan untuk mempercepat pelatihan model pada dataset gambar apusan darah. Snapshot ensemble adalah teknik deep learning untuk meningkatkan performa dan stabilitas model selama pelatihan. Sedangkan gradient camera digunakan untuk memberikan visualisasi hasil prediksi terhadap staf medis. Dengan adanya aplikasi ini, akan dapat membantu tenaga medis mempercepat dan meningkatkan akurasi selama mendiagnosis pasien. Pada penerapannya, penulis akan menggunakan bantuan library streamlit untuk pembuatan website yang akan digunakan untuk upload sekaligus resize gambar yang akan diprediksi oleh model. Hasil f1-score tertinggi yang didapatkan yaitu 99,27% dari model Ensembled EfficientNetB0 dengan menggabungkan metode transfer learning dan snapshot ensemble. Hasil tersebut didapatkan setelah melakukan pengujian model terhadap data testing yang terdiri dari 4134 gambar.
==================================================================================================================================
Malaria is a serious global health problem causing significant morbidity and mortality. Accurate and timely diagnosis of malaria is essential for effective treatment and control of the disease. In this study, the authors explored several deep learning approaches to detect the plasmodium parasite that causes malaria. The dataset to be used is the Thin Blood Smears Dataset provided by the U.S National Library of Medicine. The author compares and evaluates the performance of transfer learning techniques and ensemble snapshots on datasets which will then be assisted by the Gradient Camera. Transfer learning is used to accelerate model training on blood smear image datasets. Ensemble snapshots are a deep learning technique to improve model performance and stability during training. Meanwhile, the gradient camera is used to visualize the prediction results for medical staff. With this application, it will be able to help medical personnel speed up and improve accuracy when diagnosing patients. In its application, the author will use the help of the Streamlit library to create a website that will be used for uploading as well as resizing images that will be predicted by the model. The highest f1-score result obtained was 99,27% from the Ensembled EfficientNetB0 model by combining the transfer learning and snapshot ensemble methods. These results were obtained after testing the model against the testing data which consisted of 4134 images.

Item Type:	Thesis (Other)
Additional Information:	RSTI 006.31 Yul a-1 2023
Uncontrolled Keywords:	parasit plasmodium, transfer learning, snapshot ensemble, gradient camera, plasmodium parasite
Subjects:	T Technology > TA Engineering (General). Civil engineering (General) > TA1637 Image processing--Digital techniques T Technology > TA Engineering (General). Civil engineering (General) > TA174 Computer-aided design.
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information Technology > 59201-(S1) Undergraduate Thesis
Depositing User:	Tri Rizky Yuliawan
Date Deposited:	02 Aug 2023 02:20
Last Modified:	06 Oct 2023 04:35
URI:	http://repository.its.ac.id/id/eprint/100173

Actions (login required)

View Item