Comparative Analysis Between Gaussian Filter & Its Modified Version for High-Resolution Image Compression

Abstract

Compressing any digital image without discarding a significant portion information is of tremendous importance due to the perceived higher value of portable and limited digital storage. This study proposes an alternative to traditional Gaussian filtering scheme by including upsampling and downsampling in the process. Higher compression ratio is always desirable if the signal-to-noise ratio (SNR) is above an acceptable margin. The filtering scheme becomes more attractive for high-resolution images, since all modern devices boast in high-resolution and those files use up most of our portable storage. In this study, better compression ratio is achieved using the proposed modified Gaussian filter against a generic version of that. This study also shows tolerable SNR against the generic one, and the reduction of SNR difference between the two methods when applied to higher-resolution images.