A Foveated Channelized Hotelling Search Model Predicts Dissociations in Human Performance in 2D and 3D images

摘要

We developed a foveated search model based on the channelized Hotelling observer (FCHO) that processeseach image in parallel with Gabor channels which center frequencies decrease with retinal eccentricity (distanceaway from the fovea). The FCHO model is based on how radiologists read 3D volumes by scrolling through2D slices. The model includes a search component that explores the image by making eye movements guidedby peripheral processing and a slice scroll component that changes the current slice for the 3D images. Wedesigned an experiment consisting of a free search in 2D and 3D noise filtered with power-law statistics matchingmammograms. Two different targets were embedded in this background: one resembling a microcalcification andanother one resembling a mass. Unlike traditional model observers (Chanellized Hotelling and non-prewhiteningmatched gilter with an eye filter), performance for the FCHO search model for both 2D and 3D search tasksshowed similar dissociations as that of human observers. Detectability of a microcalcification-like target wassignificantly higher than for a mass-like target in 2D search but comparable or lower in 3D search. Together, ourresults suggest that 3D search tasks will require more computationally complex 3D search models that take intoaccount the foveated properties of the visual system and eye movements to predict human observer performancein clinically relevant tasks.