Abstract
Purpose: When the image of an object changes in size, the change may be due to either a change in object size, or the motion of an object in depth. This ambiguity may be resolved by making assumptions about the nature of motion in the world. Although many such assumptions have been suggested, consistent psychophysical support has been scarce. Methods: We present an experiment aimed at establishing the assumptions made by observers in the interpretation of object motion. Observers were presented with a simulation of an object moving around an orbit in depth. The path of this orbit was specified only by changes in image size and lateral position. We varied the extent of the orbit in depth and the function describing the change in image size in time. Three functions were used to describe the change in image size. These were analogous to a constant rate of change of image size, a constant speed along the path of the orbit, and an interpolated function halfway between the two other functions. Observers were asked whether the orbit was greater in depth or width. Results: Psychometric functions were obtained by varying the depth extent of the orbit. Three such functions were obtained for each of the size change functions. Effects of varying the size change function were examined by comparing the PSEs for each curve. The perceived extent of the orbit was most veridical when the change in image size corresponded to a constant speed along the orbit. When the rate of change of image size was constant, the orbit was perceived to be flatter. Conclusion: Our results are consistent with a mechanism biased toward minimal acceleration. That is, the visual system may assume minimal acceleration in the object and interpret changes in image size in line with this assumption. We discuss the results of this experiment with regard to Bayesian models of perception and previously reported motion constraints.
This research was funded by the EPSRC Grant No. GR/R57157/01.