作者: Benjamin W. Tatler
DOI: 10.1007/978-3-319-02868-2_2
关键词: Foveal 、 Human eye 、 Eye movement 、 Fixation (visual) 、 Visual angle 、 Computer vision 、 Computer science 、 Retinal image 、 Artificial intelligence
摘要: The manner in which we sample visual information from the world is constrained by spatial and temporal sampling limits of human eye. High acuity vision restricted to small central foveal region retina, limited just a few degrees angle extent. Moreover, effectively when retinal image relatively stabilised for periods fixation (Erdmann Dodge 1898) , last on average around 200–400 ms viewing text, scenes or real environments (Land Tatler 2009; Rayner 1998) . It clear these severe spatiotemporal constraints that high scarce resource and, like any resource, it must be distributed carefully appropriately current situation.