Parallel Processing of DCT on GPU

The Graphic Processing Unit (GPU) is increasingly becoming an important alternative for many applications that requires real time parallel processing. More interestingly, digital image processing applications such as image compression are becoming closer than ever of being processed in real time. In this paper we explore the implementation of DCT on the GPUs. Our study indicates a clear superiority of the GPU as parallel processor for image compression using DCT over the CPU. Our study also indicates that the increase in image size has considerably slows the CPU while the GPU was not affected.