AdGazer is an innovative AI model developed by researchers from the University of Maryland and Tilburg University that predicts human ad attention using eye-tracking data. By analyzing both the ad and its surrounding webpage content, AdGazer accurately forecasts how long viewers will focus on an advertisement and its brand logo. The tool demonstrates a high correlation (0.83) between its predictions and actual gaze patterns, showcasing the importance of page context, which influences one-third of ad attention. Unlike conventional ad tools that focus solely on the advertisement, AdGazer takes a holistic approach by examining the relationship between the ad and its context. With a user-friendly interface, it allows users to upload ads for gaze time predictions and heatmaps. While currently an academic tool, its architecture is poised for quick commercialization in the ad-tech space, offering marketers valuable insights into optimizing ad performance for enhanced viewer engagement.
Source link
