Cross-dataset performance evaluation of deep learning distracted driver detection algorithms

Deep learning has gained traction due its supremacy in terms of accuracy and ability to automatically learn features from input data. However, deep algorithms can sometimes be flawed many factors such as training dataset, parameters, choice algorithms. Few studies have evaluated the robustness distracted driver detection The evaluate on a single dataset do not consider cross-dataset performance. A problem arises because performance often implies model generalisation ability. Deploying real world without knowing could lead catastrophic events. paper investigates Experimental results found reveal that generalise well unknown datasets for CNN models use whole image prediction. evaluations shed light future research developing robust