Embedded DCT coding with significance masking

We evaluate two schemes for significance switching of DCT coefficients for block-based embedded DCT image compression. Both schemes deliver their best compression at any PSNR or vice versa, within a data stream that can be terminated within a few bits of any point. An image is partitioned into equal sized blocks, and a fixed point DCT of each block is calculated. The coefficients are then passed through successive 'significance sweeps' of the whole image from the most significant down to the least significant coefficient bitplanes. The coded data stream includes bits to refine only the significant coefficients at each sweep. With each new sweep, newly significant coefficients may appear within a block, and the two switching schemes evaluated are efficient methods based on block addressing and block masking. Both methods give good compression when used losslessly to the fixed point DCT precision. The best coder outperforms the baseline JPEG method in PSNR at any compression, and is similar to state of the art wavelet coders. 1. BACKGROUND The JPEG baseline method for still image coding uses the Discrete Cosine Transform (DCT) in a fixed 8x8 pixel partition [1, 2]. Through a linear quantization table and zigzag scanning of DCT coefficients, the redundancy and bandwidth characteristics of the DCT are exploited over a range of compressions. Recently, it has become clear that the JPEG coder is not effective at higher compression ratios compared to other methods such as wavelets, which have produced better results while having the advantage of being fully embedded [3, 4]. Recent success in wavelet coding relies on exploiting the structure of the Discrete Wavelet Transform (DWT) through significance switching in a tree structure. This success is a result of good structuring and quantization of the data stream, and not necessarily due to any inherent superiority of the DWT. A new round of standardization now underway, JPEG 2000, calls for a single still image coder to produce lossless and lossy compression, while being fully embedded and giving improved compression performance. Monro and Sherlock [5] have shown how DCT coding can be extended to higher compressions by relaxing the JPEG restriction on block size. This paper shows how significance switching can be applied to achieve highly accurate embedded DCT coding. By applying significance switching in combination with the zigzag scanning normally used to structure DCT data streams, the efficiency of block based DCT compression is improved at any compression …