An Ultra-Fast, Optimized and Massively-Parallelized Curvelet Transform Algorithm on GP-GPUs

The Curvelet transform is among one of the most powerful time-frequency representations of an image. However, since it is not a fast algorithm it cannot be employed in mostreal-time and/or large scale applications. This paper proposes a novel algorithm to speed up the Curvelet transform by bothoptimizing it for repetitive Curvelet usage and by massive parallelization. In this parallel algorithm we have reconfigured the structure of the Curvelet algorithm in a parallel format suitable for implementation on the GP-GPU. Experimental results prove the proposed algorithm to be ultra-fast without loss of accuracy

Keywords:

Parallel Processing , Curvelet Transform , GPU