A neural net based color recognition system

Color recognition systems currently in use today are generally microprocessor based containing a color database. Such systems, to date, are rather inflexible in their nature. As color recognition is a subset of pattern recognition, a system that is able to classify patterns would prove somewhat superior to those that use some form of sequential logic to differentiate between different colors. This study describes a new design using a combination of an artificial neural network (to replace the database) and an `expert system´ to interpret the color. The new design was seen to overcome a number of problems inherent in the microprocessor based system