PENGENALAN WAJAH DENGAN METODE CONVOLUTIONAL NEURAL NETWORK (CNN) PADA CITRA WAJAH BERMASKER

Muhammad Alif Raihan, . (2021) PENGENALAN WAJAH DENGAN METODE CONVOLUTIONAL NEURAL NETWORK (CNN) PADA CITRA WAJAH BERMASKER. Skripsi thesis, Universitas Pembangunan Nasional Veteran Jakarta.

	Text ABSTRAK.pdf Download (235kB)
	Text AWAL.pdf Download (1MB)
	Text BAB I.pdf Download (218kB)
	Text BAB II.pdf Restricted to Repository UPNVJ Only Download (1MB)
	Text BAB III.pdf Restricted to Repository UPNVJ Only Download (498kB)
	Text BAB IV.pdf Restricted to Repository UPNVJ Only Download (2MB)
	Text BAB V.pdf Download (207kB)
	Text DAFTAR PUSTAKA.pdf Download (212kB)
	Text RIWAYAT HIDUP.pdf Restricted to Repository UPNVJ Only Download (136kB)
	Text LAMPIRAN.pdf Restricted to Repository UPNVJ Only Download (4MB)
	Text ARTIKEL KI.pdf Restricted to Repository staff only Download (1MB)

Abstract

In epidemic situations such as the novel coronavirus disease (COVID-19) pandemic, face masks have become an important part of daily routine life, no exception for those who work from office (WFO). As the coronavirus COVID-19 spreads through contacts, security and attendance systems that previously used fingerprint-based or contact-based become no longer safe for users. Compared to other popular biometrics such as fingerprints, irises, palms, and veins, the face has much better potential to recognize identity in a non-intrusive manner. Hence, facial recognition is widely used in many domain applications such as surveillance, forensics, and border control. In this study, a deep learning model with the Convolutional Neural Network (CNN) method was successfully implemented for face recognition using data taken face-to-face, that is 12 subjects. The data that has been collected is preprocessed in the form of cropping, artificial face mask augmentation, resizing, and image augmentation in general. With the predetermined modeling hyperparameter configuration, the CNN models that have been built are 15 LeNet-5 models and 48 MobileNetV2 models, which were trained on Google Collaboratory cloud GPU and laptop GPU with 60:40 data split and 50 epochs of training. These models were tested on data with face mask that did not go through both stages of augmentation, which is separate from the 60:40 data. The test results are measured with accuracy for the classification of 12 classes. The highest accuracy for the LeNet-5 model is 98.15%, while for the MobileNetV2 model it is 97.22%.

Item Type:	Thesis (Skripsi)
Additional Information:	[No.Panggil: 1710511053] [Pembimbing 1: Jayanta] [Pembimbing 2: Mayanda Mega Santoni] [Penguji 1: Ermatita] [Penguji 2: Nurul Chamidah]
Uncontrolled Keywords:	face mask, image augmentation, CNN
Subjects:	Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Divisions:	Fakultas Ilmu Komputer > Program Studi Informatika (S1)
Depositing User:	Muhammad Alif Raihan
Date Deposited:	04 Nov 2021 07:31
Last Modified:	04 Nov 2021 07:31
URI:	http://repository.upnvj.ac.id/id/eprint/12015

Actions (login required)

View Item