2D CNN gabor filter FPGA implementation

Cellular neural network model was first introduced in 1988 by L.O. Chua and L. Yang. CNNhas local connections as Human?s brain has and this important makes VLSI implementationpossible. Analog constructed CNN has powerfull speeds compared to digital systems.Because of this, CNN is prefered mostly for different purposes. CNN has regular gridintercannections that have neingbourhood. The mentioned construction is convenient for mostapplications but the mostly choosen area is image processing CNN uses some filters for imageprocessing an done of them is Gabor filters.Gabor filters is used for pre processing in image processing. When implementing Gaborfilters on digital system, we match some problems due to complex manipulation. I describehow to implement Gabor fitler on digital systems with improvments, and show applicationsabout thjis implementations in my thesis. B. Shi has developed CNN Gabor filters to reducecomplex computation. This system i,s built analog VLSI and alse has the capabilty ofcomputing both real parts and imaginary parts at the same time which lead the system to anincredible speed of processing. I used some approximation methodologies to convert analogsystem to digital system in manner of implementation, then I perform CNN Gabor filteringdigitally in MATLAB and Modelsim. In addition to all, I implement this algorithm on FPGA.Key words: Cellular Neural Network (CNN), Gabor, filter, VHDL, FPGAii