Efficient use of graphics cards in implementation of parallel image processing algorithms

摘要

In this paper we present discussion of our results of parallel programming techniques for image processing applications using graphics processor units (GPU's). We have considered and analyzed various GPU's taking their standing development into account and have implemented a number of most typical image processing tasks and algorithms. On this basis we have not only estimated the capacity growth of GPU's but also the corresponding growth of performance especially for the parallel image processing.