Applications of digital signal processing, fuzzy logic, and neural nets to computational color

The matrix-vector approach to computational color allows easy application of the theory of digital signal processing (DSP). The applications demonstrated to date include color correction, color filter design, and improved interpolation and extrapolation methods. Combined with the power of new commercial software, higher computational speed of personal computers, and the availability of regional supercomputing centers, practical optimization problems involving nonlinear transformations such as CIELab can be solved in a short time. In addition to more common DSP methods, results in fuzzy logic and neural nets which have been applied to signal processing and control problems, can also be easily applied to computational color. Fuzzy set theory allows the natural inclusion of uncertainty in a color problem, for example the variation of individual observers from the CIE standard observer. Neural nets can be used to characterize processes for which there are no adequate mathematical models. These capabilities are of particular interest in the determination of the visual effects of the new fluorescent lamps which have recently been introduced.