Image Compression and Watermark Based on the Wavelet Transform

• Main Author: 廖慶賢
• Other Authors: 陳志明
• Format: Others
• Language: zh-TW
• Published: 2002
• Online Access: http://ndltd.ncl.edu.tw/handle/41019837822303159442

碩士 === 國立臺灣科技大學 === 電機工程系 === 90 === With the rapid development of Internet, The digital data can be easily access from the Internet without the permission of the original authors. For a great deal of digital data will take most of bandwidth, users have to wait for a longer period to access the data through network. Hence, there exist several very important issues for the protection of intellectual property rights and image compression. We propose an adaptive clustered zero-block algorithm for image compression in this thesis. The algorithm makes wavelet transform for each picture and uses progressive image coding. It decomposes the images into several subbands by wavelet transform, and encodes many coefficients in the same subband as clustered zero-block, if all coefficients in this block are smaller than threshold. Encoding many coefficients as a clustered zero-block symbol can achieve compression efficiently. We also introduce other compression algorithms and compare with them. The experiment result shows that proposed algorithm has superior compression performance. The wavelet-based watermarking approaches are presented in this thesis. We decompose the images into several subbands by wavelet packet transform, then find the significant coefficients in each subband and permute watermark randomly in order to achieve the encryption. Also, using a watermark self-modify scheme and image orientation to robust the watermark. The experimental results show that the proposed technique survives image processing, noise corruption, JPEG compression, linear geometry transform (rotation, shift), etc.