Visual Impaired Assistance for Object and Distance Detection Using Convolutional Neural Networks

Visual Impaired Assistance for Object and Distance Detection Using Convolutional Neural Networks

model; machine learning; vision

Vision is a very valuable giftfrom God, most aspects of human needs in the body are dominated by vision. Based on data from the World Health Organization (WHO) there are around 180 millionpeople in the world experiencing visual impairment, while the prevalence of blindness in Indonesia reaches 3 million people (1.5% of Indonesia's population), so we designed a system in the form of a prototype that could detect objects around the user and convey data in the form of sound to the user. This research discusses the application of a machine learning model using the Convolutional Neural Network method so that it can detect objects optimally. The objects that have been collected will be trained on machine learning and produce a model to be embedded in the system's main machine, namely the Raspberry PI 4B.Machine learning model training was carried out several times by changing several layer compositions until a model with optimal accuracy was obtained, however,thesize of theresulting model was quite large so the researchers carried out SSDMobileNetV2 transfer learning to obtain the optimal model.The optimal model was obtained with a model accuracy of 92% and a model size of 18 MB. Object detection testing carried out in 3 test conditions resulted in an average object detection accuracy of 84.3%, and distance detection testing carried out in 10 conditions resulted in an average distance detection error of 2.1 cm.TheResults show that the system was accurate and effective

Jumadi Mabe Parenreng1, Andi Baso Kaswar2, Ibnu Fikrie Syahputra3

https://jurnal.iaii.or.id/index.php/RESTI/article/view/5491/889

3Informatic Engineering and Computer Education, State University of Makassar, Makassar, South of Sulawesi, Indonesia

20-01-2024

FAJAR BAGUS W

PDF

ENGLIAH

TEXT