Predictable Scanpath for Gaze Points Using Classification Models

As the field of eye-tracking research advances, understanding how visual attention is allocated in dynamic environments has become increasingly relevant. This article explores the classification of scanpaths for fixations based on pupil measures, particularly in response to dynamic visual content. By examining the relationship between pupil dynamics and eye movement patterns, the aim is to enhance our understanding of visual cognition and improve applications, the novel method here is based on attention and performance applying the Euclidean path for shortest distance (EPSD) and a Region Convolutional Network embedded Support vector machine (RCNN-RCNN-SVM) for a notable correlation between pupil sizes (Constriction and Dilation) to task performance in relation with scanpaths. The result shows that people who displayed larger pupil diameters during fixations on complex areas of a webpage often also demonstrated slower task completion times and shorter scanpath, suggesting deep contemplation or struggle to process information effectively. These methods applied on scanpaths for fixations through pupil measures provide a deeper understanding of how users interact with dynamic webpages with explicit animated content. These findings can significantly aid web developers and designers in creating more intuitive and engaging interfaces that accommodate the ways users naturally navigate digital content.