IMPLEMENTATION IN REAL TIME OF DISTRACTED DRIVER DETECTION USING USING MACHINE LEARNING

IMPLEMENTATION IN REAL TIME OF DISTRACTED DRIVER DETECTION USING USING MACHINE LEARNING

Sourabh .S. Yerewar, Sarang Rohankar, Rohan Ingle, Rohan.N.Shembekar, Parag B Shahare, Prof. S.S.Ganorkar

Abstract:

In India, the death rate due to traffic accidents is one of the most serious issues. Driver distraction is a factor in about 80% of accidents. We work to create a reliable system for identifying driver distraction. There are various ways to find it, but they take longer. Early distraction detection and communication with the driver are crucial. In order to minimise the danger of accidents and detect this distraction as early as feasible, computer-based applications must be created. The goal is to create a quick and effective distraction detection tool. Some methods that have been proposed include the preprocessing, feature extraction, and classification steps. We employ a two-layer neural network to improve the result's accuracy. ELM, Softmax, and SVM-ELM are used for feature extraction and classification, respectively. The proposed method's accuracy is 97.2%, demonstrating its dependability.

Index Terms - Driver distraction, SVM , CNN, classification, machine learning.