Real-time Distracted Driver Posture Classification

TL;DR

This paper introduces a new distracted driver posture dataset and a CNN ensemble system that achieves high accuracy in real-time classification, improving safety monitoring.

Contribution

It presents a novel genetically-weighted CNN ensemble for distracted driver posture classification and evaluates the impact of visual elements on detection accuracy.

Findings

Achieved 95.98% classification accuracy with the ensemble.

Developed a real-time capable thinned ensemble with 94.29% accuracy.

Demonstrated the effectiveness of face and hand localization in distraction detection.

Abstract

In this paper, we present a new dataset for "distracted driver" posture estimation. In addition, we propose a novel system that achieves 95.98% driving posture estimation classification accuracy. The system consists of a genetically-weighted ensemble of Convolutional Neural Networks (CNNs). We show that a weighted ensemble of classifiers using a genetic algorithm yields in better classification confidence. We also study the effect of different visual elements (i.e. hands and face) in distraction detection and classification by means of face and hand localizations. Finally, we present a thinned version of our ensemble that could achieve a 94.29% classification accuracy and operate in a realtime environment.