Memory efficient self guided image filtering

Abstract

In this paper we propose a memory efficient architecture for implementation of Guided Image Filter (GIF) in Field Programmable Gate Array (FPGA) when input and guide images are the same. To reduce memory requirement we propose to use intermediate coefficients (a and b in the algorithm) directly without taking the mean to generate the filtered output. As bit widths of these coefficients are very high, the memory requirement is reduced by 66.67% while maintaining the quality of output image.