high performance computing on graphics processing units: hgpu.org

In this paper, the parallel algorithm of JPEG coding based on GPU is proposed, most image compression systems have efficiency problem and the real-time of wireless multimedia sensor networks (WMSN) which used in image compression and transmission is also an issue need to be solved, so in this paper parallel computation is used in JPEG coding, it is an effective ways to solve these problems. The system of JPEG coding system mainly has eight parts: discrete cosine transform (DCT) and inverse DCT, quantization and inverse quantization, Zig-zag ordering and inverse Zig-zag ordering, Huffman coding and decoding. The proposed parallel algorithm of JPEG coding makes all the parts of JPEG system run on GPU, so the speed of JPEG coding is improved significantly. DCT and Huffman coding are wildly used in image processing; therefore the proposed parallel algorithm can be used in many fields about image compression and processing. We use the CUDA toolkit based on GPU which is released by NVIDIA to design the parallel algorithm of DCT-JPEG algorithm. The experimental results show that compared with conventional JPEG coding, the maximum speedup ratio of parallel algorithm of JPEG coding can reach more than 120 times, and the reconstructed image has almost the same performance with the serial algorithm in terms of objective quality and subjective effect.