Energy Efficient 2D FIR Filter and Multiplier Design for Image Processing Applications

Abstract

The 2D FIR filters are simple to design which are more frequently used many more real world applications. The image processing application like medical diagnosis, pattern recognition and robot vision. This paper introduces an energy-efficient 2D Finite Impulse Response (FIR) filter design for image processing tasks like image denoising, image enhancing, image restoration etc., The 2D FIR filter is designed using transpose direct form architecture by 3x3, 5x5 and 7x7 Kernal mapping. The multiplier in this filter is replaced with the multiplier is designed using Booth-Dadda technique (which reduces number of partial products and number of additions operations). The multiplier is designed using 4:2 compressor to reduce power consumption and delay. Xilinx ISE 14.7 software with Spartan3E XC3S100E device is utilized to implement the filter in Verilog HDL and to verify the design parameters of the filter.