Skripsi

SEGMENTASI RUANG JANTUNG JANIN MENGGUNAKAN METODE YOLACT.

---

Penilaian

---

ABSTRACT Ultrasonography (USG) is a non-invasive medical procedure that uses high-frequency sound waves to produce pictures of the internal organs of the human body. This procedure can be used to examine various parts of the body, including the fetal heart. One technique for ultrasound examination of the fetal heart is through a four-chamber view which can help identify abnormalities in the fetal heart early on. This method uses Convolutional Neural Network (CNN) technology to classify the four points of view of the fetal heart and assists in determining the diagnosis and treatment of abnormalities found. In the medical world, the condition of the heart can be determined by analyzing ultrasound results by looking at the condition of the heart Image in the atria and ventricles. The method used in this study uses YOLACT and YOLACT++. There are 3 experiments used in this study, namely Abnormal, Abnormal and Normal Cases, and Abnormal with Hole Only labels with parameter values that include Epoch, Batch Size, and Learning Rate resulting in a total of 39 models. The first case produces the highest evaluation on the Darknet53 backbone with a mAP All value of 99.17% Box and 98.93% of Mask. The second case produces the highest evaluation on the Resnet50 backbone with a value of mAP All of Box 99.31% and Mask of 98.26%. And in the third case it produces the highest evaluation on the Resnet101 backbone using 500 Epoch with a mAP All value of 98.31% and a Mask of 93.44%. Apart from that, this study also made a comparison using another method, namely YOLOv7 with a Box value of 98.4% and a Mask of 98%, while YOLOv8 produced the highest Box and Mask values of 99% and 98.2%. Keywords : Instance Segmentation, USG, YOLACT, Convolutional Neural Network (CNN)

---

Availability

Inventory Code	Barcode	Call Number	Location	Status
2307002621	T114128	T1141282023	Central Library (Referens)	Available

Detail Information

Series Title

-

Call Number

T1141282023

Publisher

Indralaya : Prodi Sistem Komputer, Fakultas Ilmu Komputer Universitas Sriwijaya., 2023

Collation

xiii, 78 hlm.; ilus.; 29 cm

Language

Indonesia

ISBN/ISSN

-

Classification

005.707

Content Type

Text

Media Type

unmediated

Carrier Type

-

Edition

-

Subject(s)

Prodi Sistem Komputer
Data dalam sistem-sistem komputer

Specific Detail Info

-

Statement of Responsibility

MURZ

Other version/related

File Attachment

Comments

---

You must be logged in to post a comment