Intermediary neurons that ferry visual information between the retina and the brain are not simply connected one-to-one with the sensory cells. Each cone and rod cell in the fovea sends signals to at least three bipolar cells, whereas in the more peripheral regions of the retina, signals from large numbers of rod cells converge to a single ganglion cell. Spatial resolution in the outer portions of the retina is compromised by having a large number of rod cells feeding a single channel, but having many sensory cells participate in capturing weak signals significantly improves the threshold sensitivity of the eye. This feature of the human eye is somewhat analogous to the consequence of binning in slow-scan CCD digital camera systems.
The complex network of excitatory and inhibitory pathways in the retina are arranged in three layers of neuronal cells that arise from a specific region of the brain during embryonic development. These circuits and feedback loops result in a combination of effects that produce edge sharpening, contrast enhancement, spatial summation, noise averaging, and other forms of signal processing, perhaps including some that have not yet been discovered. In human vision, a significant degree of image processing takes place in the brain, but the retina itself also is involved in a wide range of processing tasks.
In another aspect of human vision known as color invariance, the color or gray value of an object does not appear to change over a wide range of luminance. In 1672, Sir Isaac Newton demonstrated color invariance in human visual sensation and provided clues for the classical theory of color perception and the nervous system. Edwin H. Land, founder of the Polaroid Corporation, proposed the Retinex theory of color vision, based on his observations of color invariance. As long as color (or a gray value) is viewed under adequate lighting, a color patch does not change its color even when the luminance of the scene is changed. In this case, a gradient of illumination across the scene does not alter the perceived color or gray-level tone of a patch. If the luminance level reaches the threshold for scotopic or twilight vision, the sensation of color vanishes. In Land’s algorithm, the lightness values of colored areas are computed, and the energy at a particular area in the scene is compared with all the other areas in the scene for that waveband. The calculations are performed three times, one for each waveband (long wave, short wave, and middle wave), and the resulting triplet of lightness values determines a position for the area in the three-dimensional color space defined by the Retinex theory.
Normal cones and pigment sensitivity enable an individual to distinguish all the different colors as well as subtle mixtures of hues. This type of normal color vision is known as trichromacy and relies upon the mutual interaction from the overlapping sensitivity ranges of all three types of photoreceptor cone. A mild color vision deficiency occurs when the pigment in one of the three cone types has a defect, and its peak sensitivity is shifted to another wavelength, producing a visual deficiency termed anomalous trichromacy, one of three broad categories of color vision defect. Dichromacy, a more severe form of color blindness, or color deficiency, occurs when one of the pigments is seriously deviant in its absorption characteristics, or the particular pigment has not been produced at all. The complete absence of color sensation, or monochromacy, is extremely rare, but individuals with total color blindness (rod monochromats) see only varying degrees of brightness, and the world appears in black, white, and shades of gray. This condition occurs only in individuals who inherit a gene for the disorder from both parents.
Dichromats can distinguish some colors, and are therefore less affected in their daily lives than monochromats, but they are usually aware that they have a problem with their color vision. Dichromacy is subdivided into three types: protanopia, deuteranopia, and tritanopia. Approximately two percent of the male population inherits one of the first two types, with the third occurring much more rarely.
When there is a loss of sensitivity by a cone receptor, but the cones are still functional, resulting color vision deficiencies are considered anomalous trichromacy, and they are categorized in a similar manner to the dichromacy types. Confusion often arises because these conditions are named similarly, but appended with a suffix derived from the term anomaly. Thus, protanomaly, and deuteranomaly produce hue recognition problems that are similar to the red-green dichromacy defects, though not as pronounced. Protanomaly is considered a "red weakness" of color vision, with red (or any color having a red component) being visualized as lighter than normal, and hues shifted toward green. A deuteranomalous individual exhibits "green weakness", and has similar difficulties in discriminating between small variations in hues falling in the red, orange, yellow, and green region of the visible spectrum. This occurs because the hues appear to be shifted toward red. In contrast, deuteranomalous individuals do not have the brightness loss defect that accompanies protanomaly. Many people with these anomalous trichromacy variants have little difficulty performing tasks that require normal color vision, and some may not even be aware that their color vision is impaired. Tritanomaly, or blue weakness, has not been reported as an inherited defect. In the few cases in which the deficiency has been identified, it is thought to have been acquired rather than inherited. Several eye diseases (such as glaucoma, which attacks the blue cones) can result in tritanomaly. Peripheral blue cone loss is most common in these diseases.