Real-Time Object Detection Using Various Yolo Algorithms with Audio Feedback

Authors

  • Naganjaneyulu Satuluri, G. Asritha, V. Mounika, R. Shalini

DOI:

https://doi.org/10.17762/msea.v71i3.217

Abstract

In Computer Vision, object recognition is a challenging application. It is used in many applications like security tracking, guiding visually impaired people, robotics, traffic signals and autonomous cars. Video analysis and image understanding have been improved through deep learning algorithms that work uniquely with different network architectures, that with the aim of identifying numerous objects from composite images. A large aerial image dataset called MS COCO was used to train YOLO algorithms. Raspberry Pi can be used as an external hardware source. The input is taken from the Pi camera when the button is clicked and then processed by NodeMCU and sent to Raspberry Pi which trains the data using Scaled-YOLOv4 algorithm which has high accuracy and high speed i.e. produces the accurate output in less time compared to any YOLO algorithm. When it detects the objects it generates output as audio which has object name. So it becomes easy for the user to understand the objects around them. This application is used for traffic signals, autonomous cars, security, blind people, robotics and many more things to detect objects around them.

Downloads

Published

2022-06-09

How to Cite

Naganjaneyulu Satuluri, G. Asritha, V. Mounika, R. Shalini. (2022). Real-Time Object Detection Using Various Yolo Algorithms with Audio Feedback. Mathematical Statistician and Engineering Applications, 71(3), 774 –. https://doi.org/10.17762/msea.v71i3.217

Issue

Section

Articles