The Attention Insight algorithm is trained on approximately 30800 images from eye-tracking studies. The image datasets we use are both open-source and proprietary. Data set statistics: (a) Participant attention duration: 4 seconds (b) Average gender distribution: 58% women and 42% of men (c) Average participant age distribution: varies from 7 years old to about 60+ years old. However, most participants fall into the 21–30 age bracket.

Attention Insight visual attention predictions are ~90% accurate for web images, and ~94% accurate for all other images. In addition, the accuracy of Attention Insight algorithm results was determined by comparing it to MIT’s data set containing images with eye tracking data. In order to compare eye tracking results with our algorithm results, Area under the curve (AUC Judd) metric was used, currently the main metric in MIT saliency benchmark

The Attention Insight attention prediction model is based on deep learning and can automatically detect visual attention shifts that can be used as a substitution for eye-tracking studies. Finally, to use the Attention Insight platform, users do not need to collect additional data since an AI algorithm uses previously recorded data from eye tracking studies and generates attention heatmaps in seconds.