Naufal, Muhammad Ghiyats (2026) Perbandingan metode You Only Look Once dan Long Short-Term memory untuk deteksi ekspresi wajah bayi. Sarjana thesis, UIN Sunan Gunung Djati Bandung.
Full text not available from this repository.Abstract
INDONESIA: Penelitian ini membandingkan antara model face recognition yang menggunakan basis spasial dan sekuensial, yang didasari perlunya pembanding dikarenakan belum adanya kesepakatan dari penelitian sebelumnya mengenai metode mana yang lebih baik untuk digunakan dalam model face recognition. Penelitian ini menggunakan YOLOv8 sebagai representatif arsitektur model berbasis spasial dan Convolutional Neural Network dengan Long short term Memory (CNN-LSTM) untuk arsitektur model sekuensial, kedua model diuji pada dataset berjumlah 3530 citra yang disusun dalam klasifikasi citra Tunggal dan rangkaian frame berurutan menggunakan metode penelitian CRISP-DM. Evaluasi dilakukan menggunakan metrik akurasi dan waktu inferensi untuk mengetahui efektifitas dan efisiensi komputasi kedua model. Hasil percobaan menunjukan kalau YOLOv8 menembus akurasi klasifikasi statis hingga 90% dengan inferensti yang lebih cepat, sedangkan CNN-LSTM lebih unggul dalam menangkap dinamika perubahan ekspresi . ENGLISH: This study aims to compare spatial and sequential face recognition models, based on the need for comparison due to the lack of consensus in previous studies regarding which method is better to use in face recognition models. This study uses YOLOv8 as a representative of a spatial-based model architecture and Convolutional Neural Network with Long Short Term Memory (CNN-LSTM) for a sequential model architecture. Both models were tested on a dataset of 3530 images arranged in a single image classification and a sequence of frames using the CRISP-DM research method. The evaluation was conducted using accuracy and inference time metrics to determine the effectiveness and computational efficiency of both models. The results showed that YOLOv8 achieved a static classification accuracy of up to 90% with faster inference, while CNN-LSTM was superior in capturing the dynamics of expression changes but had the disadvantage of increased computational load.
| Item Type: | Thesis (Sarjana) |
|---|---|
| Uncontrolled Keywords: | Face Recognition; Spatial; Sequential; YOLOv8; CNN-LSTM; Method Comparison; CRISP-DM |
| Subjects: | Technology, Applied Sciences |
| Divisions: | Fakultas Sains dan Teknologi > Program Studi Teknik Informatika |
| Depositing User: | Muhammad ghiyats Naufal |
| Date Deposited: | 16 Mar 2026 03:52 |
| Last Modified: | 16 Mar 2026 03:52 |
| URI: | https://digilib.uinsgd.ac.id/id/eprint/129155 |
Actions (login required)
 |
View Item |
