Models of human foveal contrast sensitivity are inadequate for the visual periphery. The main reason stems from differences in the distribution of cones and consequently, the classes of ganglion cells across the retina. In 1999, Barten adapted his parametric model of foveal contrast sensitivity to peripheral vision by adjusting for the density of ON M-ganglion cells. His influential work is improved by Bozorgian et al, in 2022, who incorporate additional physiological observations. First, they replaced the parameter of photoreceptor density with cone radius because density as an approximation for sampling resolution fails in the periphery. Next, the authors also included the ON and OFF P-ganglion cells contribution to spatial resolution. The final improvement modeled the neural noise as a function of cortical magnification rather than that based on M-ganglion cell density. This updated model reproduces aliasing effects in peripheral vision that match empirical measurements of human detection limits in the periphery. It is far superior to the original in predicting contrast sensitivity ranging from the fovea to 30 degree eccentricity with many potential applications in the realms of basic and applied vision science. A pertinent example includes designing better image compression algorithms in future gaze-contingent video technology.
You must log in
to add comments.