SISTEM DETEKSI OBJEK UNTUK PENYANDANG TUNANETRA SECARA REAL-TIME MENGGUNAKAN METODE MOBILENET BERBASIS DEEP LEARNING

RACHMANDA, RIDHO WEEDY and Fachrurrozi, Muhammad (2024) SISTEM DETEKSI OBJEK UNTUK PENYANDANG TUNANETRA SECARA REAL-TIME MENGGUNAKAN METODE MOBILENET BERBASIS DEEP LEARNING. Undergraduate thesis, Sriwijaya University.

	Text RAMA_55201_09021282025096.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (2MB) | Request a copy
	Text RAMA_55201_09021282025096_TURNITIN.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (6MB) | Request a copy
	Text RAMA_55201_09021282025096_0222058001_01_front_ref.pdf - Accepted Version Available under License Creative Commons Public Domain Dedication. Download (1MB)
	Text RAMA_55201_09021282025096_0222058001_02.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (361kB) | Request a copy
	Text RAMA_55201_09021282025096_0222058001_03.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (304kB) | Request a copy
	Text RAMA_55201_09021282025096_0222058001_04.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (348kB) | Request a copy
	Text RAMA_55201_09021282025096_0222058001_05.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (527kB) | Request a copy
	Text RAMA_55201_09021282025096_0222058001_06.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (37kB) | Request a copy
	Text RAMA_55201_09021282025096_0222058001_07_ref.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (124kB) | Request a copy
	Text RAMA_55201_09021282025096_0222058001_08_lamp.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (114kB) | Request a copy

Abstract

Visually impaired individuals facing obstacles while moving, especially on sidewalks, may experience challenges in their communication and interaction with their surroundings. Despite the frequent use of external aids such as human assistance, trained guide dogs, and canses, limitations in availability and dependence on such assistance pose constraints. This research develops a mobile object detection software system using the MobileNet method as a tool for visually impaired individuals. The mobile application aims to support their mobility on sidewalks by detecting obstacle objects and providing audiotory information to the users. Detected objects include various elements such as park benches, rocks, holes, cars, motorcycles, trees, flower pots, and streetlights. The study involves analyzing laboratory test results and direct user testing on sidewalks with the participation of visually impaired individuals in the city of Palembang. The testing process is conducted in real-time using a smartphone camera. This research conducted experiments with various batch size configurations of 8, 16, and 32 integrated with a total of 16000 steps. After conducting a series of tests, the best result was obtained by the model with a batch size configuration of 32. This result successfully achieved a classification accuracy of up to 76%.

Item Type:	Thesis (Undergraduate)
Uncontrolled Keywords:	Deteksi Objek, MobileNet-SSD, Individu Tunanetra, Aksesibilitas Trotoar, Mobilitas.
Subjects:	T Technology > T Technology (General) > T1-995 Technology (General) T Technology > T Technology (General) > T173.2-174.5 Technological change
Divisions:	09-Faculty of Computer Science > 55201-Informatics (S1)
Depositing User:	Ridho Weedy Rachmanda
Date Deposited:	16 Jul 2024 08:25
Last Modified:	16 Jul 2024 08:25
URI:	http://repository.unsri.ac.id/id/eprint/151174

Actions (login required)

View Item