Bit-plane error recovery of low frequency subbands in JPEG2000

Digital image storage and transmission is widely used in consumer products, such as digital cameras and cellular phones, where low bit-rate coding is required. In any low bit-rate encoder, such as the JPEG2000 standard, the data truncation during the encoding process, and data loss during transmission will result in lost bit-planes, which will be normally replaced by zeros at the decoder. In this paper a new algorithm has been proposed that is able to recover the lost/truncated lower bit-planes of transform coefficients in the LL subband of a wavelet transform in a JPEG2000 stream. In the proposed algorithm, the lost bit-planes of wavelet coefficients are recovered according to their distance to edges in four directions. Edges are defined based on the available bit-planes at the decoder. Simulation results indicate that the proposed algorithm achieves up to 3.25 dB improvement with respect to the simple zero-filling data recovery method. Compared to half-filling, another popular error recovery method in JPEG2000, the proposed method achieves a %60 improvement in terms of PSNR.