Skripsi

REKOGNISI WAJAH PADA CITRA BERESOLUSI RENDAH MENGGUNAKAN ARCFACE DENGAN RESTORASI BERBASIS GFPGAN

---

This study addresses the challenge of face recognition in low-resolution images captured by Unmanned Aerial Vehicles (UAVs). We propose an integrated pipeline that utilizes Generative Facial Prior-Generative Adversarial Network (GFPGAN) for image restoration and ArcFace as a feature extractor. The system was evaluated on the DroneFace dataset across varying distances and heights, comparing Cosine Similarity, SVM, and KNN as classification methods. Experimental results show a significant improvement in image quality, demonstrated by a 90.22% reduction in BRISQUE scores and a 37.52% reduction in NIQE scores. In terms of recognition performance, the Cosine Similarity method achieved the best stability with an average F1-Score increase of 10.50%. The most notable improvement was observed in long-distance scenarios (>12 meters), where the F1-Score increased substantially from 11.03% to 44.92%. However, the restoration process resulted in lower accuracy for close-range images (

---

Availability

Detail Information

Series Title

-

Call Number

T1923372026

Publisher

Indralaya : Prodi Teknik Informatika, Fakultas Ilmu Komputer Universitas Sriwijaya., 2026

Collation

xvii, 131 hlm.; ilus.; tab.; 29 cm.

Language

Indonesia

ISBN/ISSN

-

Classification

006.420 7

Content Type

Text

Media Type

-

Carrier Type

-

Edition

-

Subject(s)

Prodi Teknik Informatika
Rekognisi Wajah

Specific Detail Info

-

Statement of Responsibility

MI

Other version/related

File Attachment

Comments

---

You must be logged in to post a comment