| dc.description.abstract | The characteristics of wavelet transform are suitable for data compression and image watermark embedding. In this dissertation, the mathematical statistics in wavelet domain are analyzed. Featured-based block-wise processing is applied to image/video compression and watermarking systems. The proposed method comprises four technologies: including the significant bit-plane clustering technique for image coding (SBPC), the selective block-wise reordering with SPIHT for video coding (SBR-SPIHT), a scalable coding method based on discrete wavelet transform-MPEG (DWT-MPEG) for video coding, and the block polarity and activity index modulation for watermark embedding.
The SBPC is applied to high-energy code blocks to enhance the energy compaction by rearranging the column positions in these code blocks. The energy compaction effect can improve the coding efficiency of JPEG2000, which results in an improvement of 6.88% bit-rate reduction at 0.1 bpp on average over JPEG2000.
The SBR-SPIHT enhances the video coding efficiency for motion-compensated residuals at very low bit-rates. The motion estimation and motion compensation schemes of H.263 are used to reduce the temporal redundancy. The residuals are then wavelet transformed. The block-mapping reorganization utilizes the wavelet zerotree relationship that jointly presents the wavelet coefficients from the lowest subband to high frequency subbands at the same spatial location, and allocates each wavelet tree with all descendents to form a wavelet block. The selective multi-layer block-wise reordering technique is then applied to those wavelet blocks that have energy higher than a threshold to enhance the energy compaction by rearranging the significant pixels in a block to the upper left corner based on the magnitude of energy. Simulation results demonstrate that SBR-SPIHT outperforms H.263 by 1.28~0.69 dB on average for various video sequences at very low bit-rates, ranging from 48 to 10 kbps.
The DWT-MPEG provides resolution-scalability such that low cost existing video codec can be used to support scalable video. In each subband, a fixed-size motion compensated MPEG coder with custom-designed quantization tables and scanning direction is employed. The simulation results show that the DWT-MPEG coding method improves the image quality over ordinary MPEG coding by 0.3 ~ 1.5 dB.
A blind watermarking method is implemented to embed watermark in wavelet domain. In this blind watermarking method, block polarity is first determined based on the number of coefficients that are larger than the median value. Then, the block activity index modulation is performed based on the XOR operation of the randomized watermark and the randomized wavelet blocks polarity. Activity index modulation is represented by the sum of absolute differences (SAD) of each coefficient to the median value. The proposed method not only survives the benchmark with non-geometric, image cropping, small degree rotation, and line removal attacks, but also achieves perceptual transparency, blind detection, and low false positive rate. | en_US |