Selective fixation control for machine vision: a survey

The author describes methods for automatically directing camera gaze to new visual targets. Recent research has shown that computational advantages may be realized for vision problems when cameras are capable of fixating on an object of interest. The emphasis is not on physical aspects of controlling camera moments, but on quantitative methods used in the selection of new targets for fixation. Because many of these methods are motivated by human visual perception, a summary of relevant research in human vision is given. This is followed by a discussion of computational models for selective fixation control