Citation

Devi PR. J. Eng. Sci. (Riyadh) 2023; 14(3): 340-347.

Copyright

(Copyright © 2023, King Saud University)

DOI

unavailable

PMID

unavailable

Abstract

In India, accidents have been a leading cause of death. The cause of more than 80% of accident-related fatalities is not the accident itself but rather the delay in providing assistance to accident victims. An accident victim might be left unattended for a very long time on highways where traffic is extremely light and moving quickly. The goal is to develop a system that can recognise an accident based on a live video feed from a CCTV camera that has been mounted on a highway. Each frame of a video should be run through a deep learning convolution neural network model that has been trained to distinguish between accident- and non-accident-related video frames. It has been demonstrated that Convolutional Neural Networks are a quick and reliable method for classifying images. For comparatively smaller datasets, CNN-based image classifiers have achieved accuracy levels of over 95% and require less pre-processing than other image classifying algorithms.

Language: en