Fast and Power Reduced RNS-Based Image Filtering in Spatial Domain

This paper presents a fast and power reduced design for digital image filtering. Contrary to the typical designs for image processing, this one is based on a non-weighted number system, residue number system (RNS). The proposed structure applies a number of spatial filters on grayscale images. All the processing is done using the RNS, which results in faster and power efficient performance than those based on the standard binary number system. This paper also presents detailed comparisons with other already published papers that use the RNS for digital image processing purposes. However, these papers used improper moduli sets, hence their filtering processes are inaccurate. Contrary to those designs, the proposed one performs the filtering process accurately. The filtered images are identical to those based on the binary number system