Skripsi

DETEKSI JANTUNG ANAK MENGGUNAKAN DEEP LEARNING

---

Penilaian

---

USG or Ultrasonography is a non-invasive medical imaging method that uses high-frequency sound waves to visualize organs, tissues, and structures in the human body, including the hearts of children. This research involves a learning process using deep learning with the Convolutional Neural Network (CNN) architecture called You Only Look Once (YOLO). A total of 11 models from YOLOv7 and YOLOv8 will be used in this research experiment. The data used consists of images of children's hearts obtained from USG examinations, which will be detected using the 11 models. The dataset contains four types of heart conditions: Normal, ASD, AVSD, and VSD. Additionally, there are eight classes of cardiac chambers to be detected in this research: Right Ventricle (RV), Aorta (A), Left Atrium (LA), Left Ventricle (LV), Hole (H), Right Atrium (RA), Pulmonary Atresia (PA), and Right Ventricular Outflow Tract (RVOT). The focus of this research is on the Hole (H) class, as it distinguishes between different cases in the dataset. The parameters used in this study include the number of Epochs, Learning Rate, and Batch Size. The final results of the research include the evaluation of the training process and the testing of unseen data, which will generate outputs such as Mean Average Precision (mAP), F1 Score, Precision, Recall, Confusion Matrix, and predictions from each tested model. Among the 11 models trained and tested using unseen data, the best-performing model is yolov8x, which achieved a total mAP of 96.9% and an accuracy of 83.3% for the Hole (H) class in the training process. In the testing process with unseen data, yolov8x achieved a total mAP of 84.1% and an accuracy of 47.4% for the Hole (H) class. This research is expected to assist in the diagnosis and treatment of cardiac abnormalities in children.

---

Availability

Inventory Code	Barcode	Call Number	Location	Status
2307005555	T122857	T1228572023	Central Library (REFERENS)	Available but not for loan - Not for Loan

Detail Information

Series Title

-

Call Number

T1228572023

Publisher

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

Collation

xi, 110 hlm.; ilus.; 29 cm

Language

Indonesia

ISBN/ISSN

-

Classification

005.11 07

Content Type

-

Media Type

-

Carrier Type

-

Edition

-

Subject(s)

Pemrograman khusus
Jurusan Sistem Komputer

Specific Detail Info

-

Statement of Responsibility

ANUG

Other version/related

File Attachment

Comments

---

You must be logged in to post a comment