A Modified Low Power Architecture for Gabor Filter

This paper proposes an algorithm that explores the space of possible architectural implementation of 2D Gabor filters. The characteristics performance of the Gabor filter is analyzed by filtering the Gaussian noise, salt and pepper noise and also compared with the performance of median filter. A zero-degree kernel is generated for the Gabor function and it is convolved with the image to be filtered. The PSNR (peak signal to noise ratio) value obtained is 30.3448. The proposed scheme is utilized for Low-power architecture design and implemented for two-dimensional Gabor filter and the power consumption of the filter obtained is 0.052Watts.