Third Eye Hand Glove Object Detection for Visually Impaired using You Only Look Once (YOLO)v4-Tiny Algorithm

Abstract

An estimate of 1.3 billion people suffers visual problems, which causes a lower quality of life because sight is recognized to be vital for humans and is core in assisting them in their day-to-day activities. The study proposes a system to develop a third eye-hand glove object detection for visually challenged people with the You Only Look Once (YOLO)v4-tiny algorithm that detects indoor objects. The system captures the image using the camera attached to the Raspberry Pi 4B will be fed to the system. The object detection process will then proceed to identify object types. Distance estimation comes afterward to calculate the distance of the identified object away from the camera, both outputted through an audio output. The system included forty (40) tests of the objects from the Common Objects in Context (COCO) dataset found indoors. The system's overall F1 score, precision, recall, and accuracy is 83.00%.