Yes. For training our model we took a couple of different datasets - a dataset from an eye-tracking device and dataset gathered by using BubbleView.