ENHANCING DISTRACTED DRIVER DETECTION WITH HUMAN BODY ACTIVITY RECOGNITION USING DEEP LEARNING

• Frank Zandamela Smart Places, CSIR, Council for Scientific and Industrial Research, Pretoria, South Africa |Department of Electrical Engineering, University of Cape Town, Cape Town, South Africa https://orcid.org/0000-0003-2201-1985
• Fred Nicolls Department of Electrical Engineering, University of Cape Town, Cape Town, South Africa https://orcid.org/0000-0002-8483-412X
• Dumisani Kunene Defence and Security, CSIR, Council for Scientific and Industrial Research, Pretoria, South Africa https://orcid.org/0000-0002-9531-850X
• Gene Stoltz Defence and Security, CSIR, Council for Scientific and Industrial Research, Pretoria, South Africa https://orcid.org/0000-0001-8500-1798

DOI:

https://doi.org/10.7166/34-4-2983

Abstract

Deep learning has become popular owing to its high accuracy and ability to learn features automatically from input data. Various approaches are proposed in the literature to detect distracted drivers. However, the performance of these algorithms is typically limited to image datasets that have a similar distribution to the training dataset, which makes it difficult to apply them in real-world scenarios. To address this issue, this paper proposes a robust approach to detecting distracted drivers, based on recognising the unique body movements involved when a driver operates a vehicle. Experimental results indicate that this method outperforms current deep learning algorithms for detecting distracted drivers, resulting in a 6% improvement in classification accuracy and a two-fold improvement in overall performance (F1 score).