Abstract
Distortions of the local spatial frequency spectrum caused by motion blur may be used by the visual system to improve motion analysis (Geisler,1999; Burr, 2000; Barlow & Olshausen, 2004). We tested this hypothesis by measuring the error of perceived motion direction of 50 pairs of dots that moved (motion signal) in the presence of 50 pairs of dots without coherent motion (background noise). By manipulating the type of background pairs we were able to produce four different types of spatial amplitude spectra. The results showed that error increased greatly when the background amplitude spectrum was similar to that of the motion signal, but when the background amplitude spectrum had an anisotropy consistent with the expected motion blur of the moving signal the error was greatly reduced. The error was reduced even more than when there was no background. Shifting the spatial frequency content away from that of the motion signal also reduced the error but not as much, even when there was an anisotropy consistent with motion blur. The orientation selective “simple cells” in V1 (Hubel & Wiesel, 1962) appear likely to contribute at least as much to determining the direction of image motion as they do to static image analysis.