Skripsi

IMAGE CAPTIONING BERBASIS DEEP LEARNING PADA CITRA JANTUNG ANAK

---

Diagnosis of congenital heart disease in children, namely Atrial Septal Defect (ASD), Atrioventricular Septal Defect (AVSD), and Ventricular Septal Defect (VSD), is hindered by the limited availability of cardiologists in interpreting echocardiography images, necessitating an automated system based on artificial intelligence. This study implements and evaluates a deep learning-based image captioning model to automatically generate medical descriptions from pediatric cardiac echocardiography images across ASD, AVSD, VSD, and Normal conditions. Five encoder-decoder architectures were examined, namely DenseNet201-LSTM, ResNet101-LSTM, DenseNet201-Transformer, ResNet101 Transformer, and EchoGPT, under six preprocessing scenarios, evaluated using Bilingual Evaluation Understudy (BLEU), Metric for Evaluation of Translation with Explicit Ordering (METEOR), and Recall-Oriented Understudy for Gisting Evaluation (ROUGE-L). Long Short-Term Memory (LSTM)-based models outperform Transformer-based models, with ResNet101-LSTM as the optimal model; word tokenization yields the best generalization on unseen data, while image flipping does not contribute significantly.

---

Availability

Detail Information

Series Title

-

Call Number

T1999602026

Publisher

Indralaya : Prodi Sistem Komputer, Fakultas Ilmu Komputer Universitas Sriwijaya., 2026

Collation

xvi, 112 hlm.; ilus.; tab.; 29 cm.

Language

Indonesia

ISBN/ISSN

-

Classification

006.310 7

Content Type

Text

Media Type

-

Carrier Type

-

Edition

-

Subject(s)

Prodi Sistem Komputer
Analisis Data Komputer--Deep Learning

Specific Detail Info

-

Statement of Responsibility

MI

Other version/related

File Attachment

Comments

---

You must be logged in to post a comment