Deteksi Distraksi Visual Pengendara Berbasis Fitur Face Mesh Menggunakan Deep Learning

Putra, Niko Christian Budi (2023) Deteksi Distraksi Visual Pengendara Berbasis Fitur Face Mesh Menggunakan Deep Learning. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6022211013-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2025. Download (13MB) | Request a copy

Abstract

Kecelakaan lalu lintas adalah kejadian yang tidak diinginkan oleh semua orang ketika melakukan perjalanan. Sayangnya berdasarkan fakta yang dikeluarkan oleh WHO pada tahun 2020 [1], kecelakaan lalu lintas masih menjadi 10 penye�bab utama kematian pada negara-negara yang belum berpendapatan tinggi, serta data dari NSHTA [2] menyebutkan 38.824 orang meninggal di jalanan U.S. Kemenhub Indonesia juga mengeluarkan data bahwa kasus kecelakaan 5 tahun terakhir selalu mencapai 100.000 lebih kasus [3]. Fakta yang telah disebutkan, tentu telah menunjukan bahwa tragedi kecelakaan masih sering terjadi. Salah satu penyebab terjadinya kecelakaan adalah Driver Distraction. Jika dikelompokan Driver Distraction dapat dibagi menjadi beberapa gagguan [4], salah satu gangguannya adalah distraksi visual [5]. Pada penelitian ini, kegiatan ditraksi visual akan dideteksi dengan menggunakan masukkan keypoints posisi mata dan domain waktu dari video. Untuk keypoints akan diambil dari face mesh menggunakan mediapipe. Lalu pendeteksian kegiatan
distraksi visual akan dicoba menggunakan deep learning yang dapat mengingat informasi dari waktu sebelumnya seperti LSTM dan GRU. Penelitian ini diharapkan dapat membantu mengembangkan sistem kegiatan distraksi visual sehingga dapat mengurangi resiko terjadinya kecelakaan.
===================================================================================================================================
Traffic accidents are events that are not wanted by everyone when traveling. Unfortunately, based on the facts released by WHO in 2020 [1], traffic accidents are still the top 10 causes of death in low-income countries, as well as data from the NSHTA [2] mentions 38,824 people died on U.S. roads. The Indonesian Ministry of Transportation also released data that in the last 5 years accident cases have always reached more than 100,000 cases [3]. Of course, the facts that have been mentioned have shown that accident tragedies still often occur. One of the causes of accidents is Driver Distraction. Driver Distraction can be divided into several distractions [4], one of the distractions is visual distraction [5]. In this research, visual distraction activities will be detected by entering the key points of eye position and time domain of the video. The key points will be taken from the face mesh using the mediapipe. Then the detection of visual distraction activities will be tried using deep learning that can remember information from previous times such as LSTM and GRU. This research is expected to help develop a system of visual distraction activities so as to reduce the risk of accidents.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Face Mesh, Distraksi Visual, Deep Learning, Face Mesh, Visual Distraction, Deep Learning
Subjects:	T Technology > T Technology (General) > T174 Technological forecasting T Technology > T Technology (General) > T57.5 Data Processing T Technology > T Technology (General) > T58.62 Decision support systems
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Electrical Engineering > 20101-(S2) Master Thesis
Depositing User:	Niko Christian Budi Putra
Date Deposited:	22 Jul 2023 13:55
Last Modified:	22 Jul 2023 13:55
URI:	http://repository.its.ac.id/id/eprint/98890

Actions (login required)

View Item