KOMBINASI SEGMENTASI PRE DAN POST IVA UNTUK PENINGKATAN KINERJA DETEKSI LESI PRA KANKER

DYAKIYYAH, ANUGRA LULU and Nurmaini, Siti (2025) KOMBINASI SEGMENTASI PRE DAN POST IVA UNTUK PENINGKATAN KINERJA DETEKSI LESI PRA KANKER. Undergraduate thesis, Sriwijaya University.

Preview	Image RAMA_56201_09011182126003_cover.jpeg - Cover Image Available under License Creative Commons Public Domain Dedication. Download (1MB) | Preview
	Text RAMA_56201_09011182126003.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (4MB) | Request a copy
	Text RAMA_56201_09011182126003_TURNITIN.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (12MB) | Request a copy
	Text RAMA_56201_09011182126003_0002085908_01_front_ref.pdf - Accepted Version Available under License Creative Commons Public Domain Dedication. Download (728kB)
	Text RAMA_56201_09011182126003_0002085908_02.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (733kB) | Request a copy
	Text RAMA_56201_09011182126003_0002085908_03.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (460kB) | Request a copy
	Text RAMA_56201_09011182126003_0002085908_04.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (3MB) | Request a copy
	Text RAMA_56201_09011182126003_0002085908_05.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (271kB) | Request a copy
	Text RAMA_56201_09011182126003_0002085908_06_ref.pdf - Bibliography Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (240kB) | Request a copy
	Text RAMA_56201_09011182126003_0002085908_07_lamp.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (346kB) | Request a copy

Abstract

Cervical cancer is one of the leading causes of death among women, particularly in developing countries. Early detection through Visual Inspection with Acetic Acid (VIA) is considered effective, but it still presents a high rate of false positives due to the limitations of visual observation. This study aims to develop and evaluate the YOLOv8 model for accurate and efficient segmentation of pre- and post-VIA images to minimize false positives. Each variant of YOLOv8 was trained using different hyperparameter configurations, specifically batch size and optimizer types. Model evaluation was conducted by measuring performance metrics and comparing each YOLOv8 variant to determine the best model. Finally, unseen testing was performed to assess the model's generalization and object detection capability on previously unseen data samples. Based on the evaluation results, the best-performing model was YOLOv8m-seg with a batch size of 8 and the Adam optimizer.

Item Type:	Thesis (Undergraduate)
Uncontrolled Keywords:	Cervical cancer, YOLOv8, Segmentation
Subjects:	Q Science > Q Science (General) > Q334-342 Computer science. Artificial intelligence. Algorithms. Robotics. Automation.
Divisions:	09-Faculty of Computer Science > 56201-Computer Systems (S1)
Depositing User:	Anugra Lulu Dyakiyyah
Date Deposited:	15 Jul 2025 06:50
Last Modified:	15 Jul 2025 06:50
URI:	http://repository.unsri.ac.id/id/eprint/178564

Actions (login required)

View Item