Skripsi

KLASIFIKASI CITRA UNTUK MENDETEKSI KELUMPUHAN WAJAH MENGGUNAKAN ARSITEKTUR VGGFACE

---

Facial paralysis is a condition characterized by the loss of motor function in the facial muscles and requires accurate clinical evaluation. Conventional visual assessments are often subjective, creating variability in diagnosis; therefore, a more consistent image-based assistance system is needed. This study develops an automated classification system for facial paralysis severity using the VGGFace Convolutional Neural Network with a Transfer Learning approach adapted to the six categories of the eFACE scale. The dataset from the Massachusetts Eye and Ear Infirmary was augmented to address limited data and class imbalance. Software development followed the Rational Unified Process and was implemented as a web-based application using Streamlit. The evaluation results show that the model configuration using the Adam optimizer with a learning rate of 0.0001 produced the best performance with an accuracy of 99.6% and stable evaluation metrics. The Stochastic Gradient Descent optimizer did not reach convergence. In external validation, the model generated 14 correct predictions out of 20 test images, indicating adequate generalization. The system achieved an average inference time of 0.5373 seconds per image, making it suitable as an early screening tool.

---

Availability

Detail Information

Series Title

-

Call Number

T1897122025

Publisher

Indralaya : Prodi Teknik Informatika, Fakultas Ilmu Komputer Universitas Sriwijaya., 2025

Collation

xv, 125 hlm.; ilus.; tab.; 29 cm

Language

Indonesia

ISBN/ISSN

-

Classification

006.370 7

Content Type

Text

Media Type

-

Carrier Type

-

Edition

-

Subject(s)

Prodi Teknik Informatika
Algoritma Deteksi Citra Wajah

Specific Detail Info

-

Statement of Responsibility

MI

Other version/related

File Attachment

Comments

---

You must be logged in to post a comment