Parallel JPEG2000 image coding on multiprocessors

Abstract

In

Energy efficient distributed image compression in resource-constrained multihop wireless networks

“…Once a node completes its 2-D wavelet transform, it sends the result back to a center node. The processing of tiles independently leads to a rate-distortion loss and blocking artifacts as the number of tiles and/or processors is increased [16]. We found that the quality loss and blocking artifacts are small when the number of tiles is small or when the bit rate of compressed image is not very low.…”

“…Thus, our proposed approach of image compression intersects with the literature on parallel wavelet transform, which primarily focuses on parallel DWT on special purpose hardware (e.g., application specific VLSI architectures [12], FPGAs [13], DSPs [14]) or general purpose multiprocessor systems [15], [16]. The difference, of course, is that previous parallel wavelet transform research assumes the energy cost of data movement between processors to be zero or very small compared to the wavelet transform computation.…”

Hardware Implementation of a Real Time Image Compression

“…Many filters for compression were introduced many masks of DWT and one important application is JPEG2000 standard [4,5,6]. DWT is performed via the implementation of low pass filter (LPF) and high pass filter (HPF) on the original data [7,8,9].…”

Optimizing JPEG2000 Still Image Encoding on the Cell Broadband Engine

“…Bit modeling and arithmetic coding are performed in Tier-1, whereas tag tree building is performed in Tier-2. Prior analysis of JPEG2000 execution time [1,8,9] shows that Tier-1 coding in the EBCOT and the DWT are the most computationally expensive algorithmic kernels.…”

“…JPEG2000 library adopts Jasper [1], a still image transcoder, as a baseline program. Jasper is previously parallelized by Meerwald et al [9] using OpenMP. However, the authors parallelize Tier-1 coding in the EBCOT and the DWT only to minimize the code modification.…”