Accelerating Haar wavelet transform with CUDA-GPU (July 2017)
Esmael Torabi Angaji, Seyed Ali Razavi Ebrahimi · 2017
Haar transform is the simplest wavelet transform that has been used for image compression by representing image in new space with orthonormal basis vectors. This algorithm has a high potential capability for a parallel implementation. In this paper we have implemented an appropriate parallel Haar transform on CUDA-GPU and a reduction of the computational time to less than one tenth of the sequential run time is achieved.