Skripsi

SEGMENTASI THORAX JANTUNG JANIN DENGAN MENGGUNAKAN YOU ONLY LOOK ONCE

---

Penilaian

---

Early detection of fetal heart anomalies, such as cardiomegaly, is critical in prenatal diagnosis. However, manual segmentation of ultrasound (USG) images requires specialized expertise and is time-consuming, highlighting the need for an efficient automated solution. This research aims to perform automatic segmentation of fetal thorax and heart areas using You Only Look Once (YOLO) version 8 and 11 models, with Nano, Small, Medium, and Large variants. The dataset comprises 762 annotated fetal thorax ultrasound images. Model performance was evaluated using mAP (mean Average Precision) and IoU (Intersection over Union) metrics on validation and unseen data. The results indicate that YOLOv11 outperforms YOLOv8, with optimal configurations as follows: YOLOv11 Nano (epoch 200, mAP@50 = 99.1%, IoU = 85.9%), YOLOv11 Small (epoch 100, mAP@50 = 98.6%, IoU = 85.5%), YOLOv11 Medium (epoch 200, mAP@50 = 98.9%, IoU = 86.1%), and YOLOv11 Large (epoch 100, mAP@50 = 99.5%, IoU = 86.5%). These results suggest that YOLOv11 models are effective and reliable for segmenting thorax and heart regions in fetal ultrasound imagery.

---

Availability

Inventory Code	Barcode	Call Number	Location	Status
2507004115	T178507	T1785072025	Central Library (REFERENCE)	Available but not for loan - Not for Loan

Detail Information

Series Title

-

Call Number

T1785072025

Publisher

Indralaya : Prodi Sistem Komputer, Fakultas Ilmu Komputer Universitas Sriwijaya., 2025

Collation

xxi, 358 hlm.; ilus.; tab.; 29 cm.

Language

Indonesia

ISBN/ISSN

-

Classification

004.07

Content Type

-

Media Type

-

Carrier Type

-

Edition

-

Subject(s)

Sistem komputer

Specific Detail Info

-

Statement of Responsibility

TUTI

Other version/related

File Attachment

Comments

---

You must be logged in to post a comment