Skripsi

PERBANDINGAN KINERJA PENGENALAN WAJAH MENGGUNAKAN METODE FACENET PYTORCH DAN KERAS FACENET

---

Penilaian

---

Face recognition has become an important technology in various applications, but challenges arise when multiple faces need to be recognized simultaneously in a single image or video frame. This research develops a multiple face recognition system using the Multi-Task Cascaded Convolutional Neural Network (MTCNN) method for face detection, Facenet models with Pytorch and Keras frameworks for face recognition, and Support Vector Machine (SVM) for classification. This research compares the performance of Facenet Pytorch and Keras Facenet in terms of processing speed, memory usage efficiency, and recognition accuracy. Using a dataset of 1000 images taken from 10 different classes with training and testing data percentages of 70%:30% and 80%:20%, this research shows that Facenet Pytorch is faster and more efficient in memory usage. The average time required by Facenet Pytorch for the embedding process is 0.15 seconds per image, while Keras Facenet requires 0.86 seconds. Facenet Pytorch also uses less RAM, 384.19 MB lower than Keras Facenet. Although Facenet Pytorch uses 3% more CPU, its speed and memory efficiency make it more suitable for applications that require fast response and low memory usage. In system testing, Facenet Pytorch outperforms Keras Facenet with an average time of 10.549 seconds faster and more efficient memory usage of 13.98 MB, despite using 64.254% more CPU. Both models can accurately recognize faces, but Facenet Pytorch generally provides higher and more consistent confidence scores. This study concludes that Facenet Pytorch is more efficient and reliable in multiple face recognition, although it requires further optimization in CPU usage. Keywords: multi-face recognition, MTCNN, Facenet, SVM, Pytorch, Keras

---

Availability

Inventory Code	Barcode	Call Number	Location	Status
2507001256	T164748	T1647482024	Central Library (Reference)	Available but not for loan - Not for Loan

Detail Information

Series Title

-

Call Number

T1647482024

Publisher

Indralaya : Prodi Magister Ilmu Komputer, Fakultas Ilmu Komputer, Universitas Sriwijaya., 2024

Collation

xiv, 95 hlm.; ilus.; tab.; 29 cm.

Language

Indonesia

ISBN/ISSN

-

Classification

004.07

Content Type

Text

Media Type

-

Carrier Type

-

Edition

-

Subject(s)

Ilmu Komputer

Specific Detail Info

-

Statement of Responsibility

MI

Other version/related

File Attachment

Comments

---

You must be logged in to post a comment