摘要
A new technique for embedding visible watermarks into a digital image isdescribed in this paper, which uses statistical properties of each pixel and its immediatesurrounding pixels in the host image, and also the statistical properties of the host image and thatof the image to be embedded (watermark), as the pixels in the host image are replaced one by one bythe pixels in the watermark image. The aim is to establish a relationship between pixels in theimages that will make the embedded watermark robust and semitransparent, and to minimize errorscaused by watermark addition and subtraction. Also, the original image can be completelyreconstructed from the watermarked images.
A new technique for embedding visible watermarks into a digital image isdescribed in this paper, which uses statistical properties of each pixel and its immediatesurrounding pixels in the host image, and also the statistical properties of the host image and thatof the image to be embedded (watermark), as the pixels in the host image are replaced one by one bythe pixels in the watermark image. The aim is to establish a relationship between pixels in theimages that will make the embedded watermark robust and semitransparent, and to minimize errorscaused by watermark addition and subtraction. Also, the original image can be completelyreconstructed from the watermarked images.
基金
theNationalNaturalScienceFoundationofChina(GrantNo .60 0 72 0 3 0 )andtheKeyDisciplinaryDevelop mentProgramofShanghai(GrantNo .2 0 0 1 44 )
