A Deep Learning Approach for Driver Monitoring System

Abstract

In recent years, the alarming rise in accidents attributed to distracted and drowsy driving has prompted a concerted effort by automotive researchers and manufacturers to implement innovative technological solutions. The proposed project uses state of the art Convolutional Brain Organizations (CNNs) to recognize basic states, working through two interconnected modules. The primary module examines ongoing pictures or video feeds to remove facial elements, critical for grasping the driver's state. These highlights act as contribution for the subsequent module, which utilizes a lightweight profound learning design enhanced for edge gadgets. This approach improves discovery precision, giving an exhaustive arrangement contrasted with past techniques utilizing multi-facet perceptrons