Double opponent cell

Figure 2.13 Four different single opponent cells are shown in (a). Two double opponent cells are shown in (b).

Neural Architecture based on Double Opponent Cells... 

Dufort and Lumsden (1991) proposed a model for color categorization and color constancy. In their model, color constancy is achieved by using the output from double opponent cells... 

Color processing in VI is confined to small circular areas, known as blobs, in which double-opponent cells are found. They display a center/surround behavior based on the red reen and yeUow-blue axes but lack orientation selectivity. The V1 blobs were first identified by their uptake of certain enzymes, and only later was their role in color vision discovered [Livingstone and Hubei, 1984]. The blobs are especially prominent in layers 2 and 3, which receive input from the P cells of the LGN. 

VI mainly connects to area V2, which surrounds VI (Tov6e 1996). Area V2 seems to be organized into three types of stripes, the so-called thick, thin, and interstripes. The stripes seem to be used to process visual orientation (thick stripes), color (thin stripes), and retinal disparity (interstripes). Adjacent stripes respond to the same region of the visual field. Neurons of layer 4B of V1 connect to the thick stripes. Cells found inside the thick stripes are selective for orientation and movement. Many of the cells also respond to retinal disparity. The neurons of the blobs are connected to the thin stripes. These cells are not orientation selective. More than half of these cells respond to color. Most show a double opponent characteristic. The cells of the interblob region connect to the interstripes. Neurons of the interstripe region respond to different orientations but neither to color nor to motion. A condition know as chromatopsia is caused by damage to certain parts of VI and V2. Individuals who suffer from chromatopsia are not able to see shape or form. However, they are still able to see colors. 

---