Beginning with Ratliff and Hartline's (1959) description of lateral inhibition in the compound eye of the horseshoe crab Limulus polyphemus and further work in this area by Ratliff et al. (1963), Ratliff et al. (1966) and Lange et al. (1966), it became well established that lateral inhibition could be interpreted in the spatial frequency domain as a linear spatial filtering operation that enhanced contours and improved resolution of the compound eye of Limulus.
Bliss and Macurdy (1961) observed that many human visual contrast phenomena, such as Mach's bands, can be described mathematically in the spatial frequency domain as linear spatial filtering operations, which presumably take place at the retinal and cortical level. Bliss and Macurdy showed how spatial impulse responses (to a point source of light) were related to the spatial filtering operations, and extended their models into discrete space by use of the z-transform.
The following section examines a nonlinear, logic-based approach to visual feature extraction system described by Zorkoczy (1966). Zorkocy models used regularly spaced arrays of binary receptors (ON or OFF) that feed into asynchronous, sequential logic circuits with simple operations such as inversion, unit delay, AND, and OR. The features of simple black and white moving objects were discriminated.
In Section 7.1.2, the application of two-dimensional, layered neural models emulating linear spatial filters based on the work of Fukushima (1969; 1970) is considered. Again, the vertebrate retina is the basis for these feature extraction models. Fukushima's models assume five layers of nodes (excluding the receptor layer). Each layer projects non-negative pulse frequency signals to adjacent layers, which are considered non-negative, continuous analog variables in describing their operations. Signals are conditioned by weightings assigned to their paths. More than one path may converge on a node, and signal paths can backpropagate.
Was this article helpful?
This guide will help millions of people understand this condition so that they can take control of their lives and make informed decisions. The ebook covers information on a vast number of different types of neuropathy. In addition, it will be a useful resource for their families, caregivers, and health care providers.