Pyramidal Image Blending Using CUDA Framework
Computer Division, Bhabha Atomic Research Centre, Trombay, Mumbai, Maharashtra, India 400085
International Journal of Engineering Science and Technology (IJEST), Volume 3, No. 12, 2011
@article{shete2011pyramidal,
title={Pyramidal Image Blending Using CUDA Framework},
author={Shete, Pritam Prakash and Venkat, P. P. K. and Bose, S. K.},
journal={International Journal of Engineering Science and Technology (IJEST)},
volume={3},
number={12},
year={2011}
}
We propose and implement a pyramidal image blending algorithm using modern programmable graphic processing units. This algorithm is an essential part of an image stitching process for a seamless panoramic mosaic. The CUDA framework is a novel GPU programming framework from NVIDIA. We realize significant acceleration in computations of the pyramidal image blending algorithm by utilizing the CUDA as a computational resource. Specifically we demonstrate the efficiency of our system by parallelization of the algorithm and optimization of the memory resources of the GPU. We compare the execution time of the CPU as well as various CUDA based implementations. Just parallelization of the algorithm by the CUDA framework provides 20 times speedup, whereas optimizing the GPU memory IO gives more than 30 times speedup.
January 5, 2012 by hgpu
