Face Recognition Attendance System (FERAS) Untuk Mendukung Program Work From Anywhere (WFA)

Nababan, David (2024) Face Recognition Attendance System (FERAS) Untuk Mendukung Program Work From Anywhere (WFA). Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5022201263-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2026. Download (7MB) | Request a copy

Abstract

Seiring berkembangnya teknologi ternyata berdampak pada budaya kerja mulai dari sistem kerja Work from Office (WFO), Work from Home (WFH), maupun hybrid. Bukan hanya karena kasus covid-19 yang mendorong kemunculan WFH, alasan kenyamanan mendorong banyak orang untuk memilih sistem kerja yang tidak terlalu diikat oleh waktu, lokasi, maupun Outfit of The Day (OOTD). Dari fenomena tersebut kemudian muncul sistem kerja dengan istilah Work from Anywhere (WFA). Tren ini memungkinkan pegawai dapat bekerja dari mana saja dan kapan saja. Namun, sistem ini bukan tidak memiliki tantangan terutama terkait dengan tingkat kedisiplinan pegawai dan sistem pengawasan yang sangat lemah. Berangkat dari permasalahan tersebut dalam tugas akhir ini menawarkan sistem presensi kerja dengan nama Face Recognition Attendance System (FERAS) dengan konsep berbasis pengenalan wajah secara real time lewat kamera laptop. Dalam proses pengembangannya, sistem FERAS dibangun menggunakan metode pengolahan sinyal multimedia dan aplikasi Deep Learning dengan Convolutional Neural Network (CNN) dan diperkaya dengan berbagai library seperti Tensorflow, dan OpenCv. Selanjutnya dilakukan pencarian training-model terbaik lewat proses training dan validasi terhadap sejumlah dataset yang diujicobakan, yaitu 35.280 citra training dan 7.560 citra validasi yang diulang sebanyak 39 iterasi (epoch) dan menghasilkan data training-model dengan tingkat akurasi 98,86% dan Loss 47,63% yang terjadi pada epoch ke-34. Selanjutnya, training-model yang diperoleh diimplementasikan pada sistem FERAS. Setelah dilakukan pengujian terhadap sistem FERAS dengan berbagai skenario maka dapat disimpulkan bahawa sistem FERAS memiliki kinerja rata-rata sebesar 91,29%.
=================================================================================================================================
As technology advances, it has significantly impacted work culture, ranging from Work from Office (WFO), Work from Home (WFH), to hybrid systems. The emergence of WFH was not only driven by the COVID-19 pandemic but also by the comfort and flexibility it offers, leading many people to prefer a work system that is less bound by time, location, or Outfit of The Day (OOTD). From this phenomenon, a new work system known as Work from Anywhere (WFA) has emerged. This trend allows employees to work from anywhere and at any time. However, this system is not without its challenges, particularly concerning employee discipline and the weakness of the supervision system. To address these issues, this proposes a work attendance system named the Face Recognition Attendance System (FERAS), which uses real-time face recognition through a laptop camera. In its development process, the FERAS system is built using multimedia signal processing methods and Deep Learning applications with Convolutional Neural Networks (CNN), enhanced with various libraries such as TensorFlow and OpenCV. Subsequently, the best training model is sought through the process of training and validation on several datasets, namely 35,280 training images and 7,560 validation images, repeated for 39 epochs, resulting in a training model with an accuracy rate of 98.86% and a loss of 47.63% occurring at epoch 34. The training model obtained is then implemented in the FERAS system. After testing the FERAS system with various scenarios, it can be concluded that the FERAS system has an average performance of 91.29%.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Convolutional Neural Network, Pengolahan Citra Digital, Rekognisi Wajah, Sistem Presensi, Work from Anywhere (WFA), Attendance System, Deep Learning, Face Recognition, Image Processing
Subjects:	T Technology > T Technology (General) > T58.5 Information technology. IT--Auditing T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5101 Telecommunication
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Electrical Engineering > 20201-(S1) Undergraduate Thesis
Depositing User:	DAVID NABABAN
Date Deposited:	29 Jul 2024 06:47
Last Modified:	29 Jul 2024 06:47
URI:	http://repository.its.ac.id/id/eprint/109629

Actions (login required)

View Item