Research indicates that a significant component of human eye movement behavior constitutes a set of consistent biases independent of visual content, the most well-known of which is the central bias. While all prior art focuses on representing saccadic motion and biases in Cartesian retinotopic coordinates, here we propose the Polar Saccadic Flow model, a novel approach for modeling saccades' space-dependent biases in a polar representation. By breaking saccades into orientation and amplitude, the Polar Saccadic Flow model enables more accurate modeling of these components, leading also to a better understanding of the saccadic bias. Moreover, the polar representation also uncovers hitherto unknown patterns and biases in eye movement data, allowing for a more detailed and nuanced analysis of saccadic behavior. These findings have implications for the study of human visual perception, can help to develop more accurate eye movement models, and also may improve eye tracking technologies.
Read full abstract