Extracting the watermark requires the original image , and its procedure is the reverse of embedding procedure . the existence of watermark is decided by observation visually and similarity measurement 水印恢复时需借助原始图像信息，其过程是水印嵌入的逆过程，水印的存在性通过视觉观察和相似性度量两种标准来判定。
Numerical experiments show that , under various image processing and attacks , the proposed algorithm has such advantages as good robustness , high security , requiring not the original image for watermarking extraction , etc 实验表明，在不同图像处理和攻击下，该算法具有鲁棒性好、安全性高、提取水印时不需要原始图像等优点。
While sub - images in the high resolution ratio are details of original image . this progress is called image wavelet decomposition . reverse wavelet transform put these sub - images together so that original image is recovered 而小波的逆变换即是将这些不同分辨率上的子图组合起来以恢复出原始图像，这个过程叫做图像的小波重构。
It is for the reason that the image can change to be more abstract and impacted after segmenting the image , separating the objects , extracting the features , measuring the parameters and so on 这是由于通过图像分割、目标分离、特征提取、参数测量等技术可以将原始图像转化为更抽象更紧凑的形式，使得更高层的图像分析和理解成为可能。
In so doing the original image is decomposed into lots of integral sub - blocks , and the position , size and gray value of each integral sub - block can be stored with an efficient way , so as to achieve the aim of compression 该方法把原始图像分解为若干个整体子块，并采用有效的方式存储每个整体子块的位里、大小和灰度，从而达到压缩的目的。
Finally , we recover the watermark according to encodement . we also apply the novel detection strategy into the image watermarking algorithm based on fourier - mellin transformation , which is trong to geometric invariants 提出了构造和提取“零水印”的算法： “零水印”是利用图像的重要特征来构造水印信息，而不修改原始图像的任何数据。
Thinking over the computer calculation speed and the requirement of low cost , and combines the system needs , puts forward the newer and quicker algorithm for studies the questions such as filtering and segmentation of original image 考虑计算机运算速度和低成本的要求，结合系统需要，提出更新更快的算法，研究原始图像的滤波、分割等问题。
For example , if you consider an image rendering grid , the data that makes up the original image description and the code functions or applications that converts this vector data into the final image will remain the same 例如，如果您考虑一个图像呈现网格，构成原始图像描述的数据以及将这些向量数据转换成最终图像的代码（函数或应用程序）都是相同的。
