JURNAL TEKNIK INFORMATIKA DAN SISTEM INFORMASI
Vol 9 No 3 (2022): JATISI (Jurnal Teknik Informatika dan Sistem Informasi)

Analysis of EfficientNetV2 Model Usage in Predicting Gender on the Face of Mask Users

Novendra Adisaputra Sinaga (Unknown)



Article Info

Publish Date
14 Sep 2022

Abstract

A technique for identifying physical traits or human behavior that is utilized as input for pattern recognition is called biometrics. Each type of biometric identification undoubtedly employs a unique technology. In order to do research on how to promote or sell items in accordance with visitor gender, a gallery or exhibition, such as a movie theater, retail mall, or exposition, needs visitor information from the event. The EfficientNetV2 model, a New Family in the Covolution Neural Network (CNN) family, outperforms the previous model in terms of parameter efficiency and training speed. According to tests, the EfficientNetV2 model can learn up to 6.8.The results using the EfficientNetV2 model were carried out for 25 epochs and there were 2 classes, namely male and female, each of which consisted of 72,318 training data and 16,813 testing data. The accuracy value for training is 0.9455 (94.5%) and for data testing the accuracy value is 0.9475 (94.7%). The loss value for training is 0.1375 (13.75%) and for testing data the loss value is 0.1277 (12.7%).

Copyrights © 2022






Journal Info

Abbrev

jatisi

Publisher

Subject

Computer Science & IT

Description

JATISI bekerja sama dengan IndoCEISS dalam pengelolaannya. IndoCEISS merupakan wadah bagi para ilmuwan, praktisi, pendidik, dan penggemar dalam bidang komputer, elektronika, dan instrumentasi yang menaruh minat untuk memajukan bidang tersebut di Indonesia. JATISI diterbitkan 2 kali dalam setahun ...