Abstract
Abstract: :
Purpose: Visual information is encoded by the photoreceptor mosaic as an array of picture elements (pixels). If the sampling frequency of the retina is lower than the optical resolution of the eye, as is the case in at least some species of fish, patterns in the image may interfere with the mosaic of cones. The resulting interference patterns – aliases – constitute incorrect visual information that is conveyed to the CNS. It was investigated how the retina of the highly visual fish species Aequidens pulcher (Cichlidae) is adapted to the problem of aliasing. Methods: A. pulcher has a trichromatic cone mosaic consisting of single short–wave–sensitive cones and pairs of middle– and long–wave–sensitive cones (double cones). The mosaic is of crystal–like regularity, with single cones at the centers of squares formed by four double cones each (square–type mosaic). The performances of pixel mosaics were investigated by computer modeling. The model included partial electrical cross–talk between the members of double cones, such as present in fish retinae. Various test patterns were sampled with various types of pixel mosaic and low–pass filtered to emulate overlap in acceptance angle between neighboring cones. The dynamic ranges of the images were normalized to take into account adaptive mechanisms in the retina. Results: The square–type cone mosaic generated fewer and weaker aliases than mosaics of hexagonal and other geometries, e.g., mosaics used in manufactured imaging devices. Partial electrical cross–talk between double cones reduced the amplitude of aliases, with little loss of spatial information. Conclusions: Aliasing may have been an important factor in the evolution of different types of cone mosaic. Of all tested mosaics, the square–type cone mosaic of many fishes, including double cones and electrical cross–talk between their members, was least likely to generate aliases and the few aliases that occured were low in amplitude. It may be possible to implement similar mosaics in manufactured imaging devices to improve visual interfacing between humans and machines.
Keywords: photoreceptors • color vision • scene perception