PARALLEL IMPLEMENTATION OF DIGITAL WATERMARKING TECHNIQUE ON GPU

Abstract

Watermarking technique is a well known technique since ages past. Invisible watermarking . technique is profoundly called to be the complex one in terms of computation because of its incorporation with intensive algorithms. Improving the efficiency of watermarking process has become one. of the challenges in image authentication today. Hence, there have been large amount of'research going on to optimize the performance of these watermarking schemes and methods. The schemes proposed till now performed well in verification results. But most of them are time consuming traditional serial manners. It's time to make it-high performable and flexible. In this work _ we present .a Compute Unified Device Architecture (CUDA) based implementation of two image authentication algorithms on Graphics Processing Unit (GPU) devices. Two unique devices having different clock rate are considered for implementation. Taking Speed Up as an attribute Graphics Processing Unit (GPU) parallel - implementation performs much better than that of CPU serial implementation.