Driver_isdrowsiness_4096.csv

The original driver drowsiness dataset contains 22,348 drowsy images and 19,445 non-drowsy images, where each image is 227x227x3 pixels, which means that the feature information is 154,587 (i.e., 227x227x3) dimensions. Here, we formed the driver_drowsiness(1) by extracting 4096-dimensional features from the original datasets using the VGG16 model. Namely, driver_drowsiness(1): 22,348 drowsy images and 19,445 non-drowsy images, 4096-dimensional features, and 2 classes.