A Study on Self-embedding Fragile Watermarking Based on Characteristics of the Average

• Main Authors: Zhang, Fang-Lun, 張芳綸
• Other Authors: Wang, Lung-Jen
• Format: Others
• Language: zh-TW
• Published: 2016
• Online Access: http://ndltd.ncl.edu.tw/handle/38055933528140081766

碩士 === 國立屏東大學 === 資訊工程學系碩士班 === 104 === Over the times, scientific and technological progress faster and faster. The speed of information transfer is also proportional to the rate of growth. There are daily tens of thousands of images circulating on the Internet, where security is very important. In order to avoid the image being tampering, we proposed a watermark technology for image restoration to defend the copyright of authors. As a start, let the host image that will be divided into 4x4 non-overlapping blocks, and each block is decomposed by discrete wavelet transform (DWT).Next the low-frequency coefficients of each DWT block is used to calculate the characteristics of the average which can be embedded as the feature watermark, as an average watermark. In addition, the characteristics of the average compared with each low-frequency coefficients which can be embedded as the feature watermark. Finally, choose MSB5 of the restore block and average watermark are used as recover watermark, respectively. And then make a complete image with inverse wavelet conversion, embedding process is completed. Image restoration way of block is if it is determined “tampered block”, then using average watermark and recover watermark to replace it. Finally, the experiment results show that, the proposed method compared with recent researches, obtains a better PSNR value and image quality after recovery of the tampered regions.