Image of PENGENALAN WAJAH PADA CITRA RESOLUSI RENDAH MENGGUNAKAN ARCFACE DAN GFPGAN
Bookmark Share

Skripsi

PENGENALAN WAJAH PADA CITRA RESOLUSI RENDAH MENGGUNAKAN ARCFACE DAN GFPGAN

Facial recognition, as a result of the integration of artificial intelligence and computer technology, continues to develop and has the potential for widespread application, but most existing methods are trained on high-resolution images in controlled environments. In real-world conditions, limitations in camera distance, device quality, and environmental conditions make it difficult to obtain high-resolution images, making low-resolution face recognition a challenge. This study aims to evaluate the performance of facial recognition models on low-resolution images using the QMUL-TinyFace dataset, and to examine the effect of super-resolution on identification performance. The restoration method used is GFP-GAN to handle complex degradation, while facial recognition is performed with ArcFace. Evaluation on the ORL dataset (a simple dataset) uses accuracy, precision, recall, and F1-Score metrics, while evaluation on the QMUL-TinyFace dataset uses the Rank-K identification metric. Rank-K testing is performed on a probe–gallery scheme, with the probe folder containing 3,513 images and the gallery folder containing 4,179 images from 2,471 identities. The test results showed the highest Rank-1 accuracy of 70.83% for restored images and 74.1% for unrestored images. This finding indicates that improving visual quality through restoration does not always improve identification performance.

Availability
#
Central Library (Reference) T1892862025
T189286
Available but not for loan - Not for Loan
Detail Information
Series Title
-
Call Number
T1892862025
Publisher
Indralaya : Prodi Teknik Informatika, Fakultas Ilmu Komputer Universitas Sriwijaya.,
Collation
xviii, 166 hlm.; ilus.; tab, 29 cm
Language
Indonesia
ISBN/ISSN
-
Classification
004.930 7
Content Type
Text
Media Type
-
Carrier Type
-
Edition
-
Subject(s)
Prodi Teknik Informatika
Proses Arcface dan Gfpgan
Specific Detail Info
-
Statement of Responsibility
Other version/related
File Attachment
Comments
You must be logged in to post a comment