Simplified Color Transformations for Analog Processing in CMOS Imagers
Faculty of Computing, Health and Science
School of Engineering and Mathematics
The paper presents an approach to color coding with the view of integration into the intelligent CMOS imager using analog processing of the photocurrent. The basic theory behind the choice of color filters is reviewed. Several transformations are analyzed and tested in the entire range of visible spectrum. The reconstructed RGB signals from the Optimized Basic Transformation (OBT) have the lowest error rate compared to other methods. It has almost 0% error in R and B channels, bnt significantly higher in the G channel. However, when testing with digital images, the error rates generated from OBT is reasonable low, 5.8% for G channel, 0.7% for Rand 1.2% for B channels.