The Research And Application Of Image Segmentation On Document Image Watermarking

With the fast development of Internet and multimedia, how to protect the information of multimedia and digital products has already been an important issue cared by all the countries in the world. Digital watermarking occurred in such a background, which can make up for the shortcomings of traditional methods of information cryptology. Digital watermarking has widely used fields in authentication, counter-counterfeiting, counter-tampering, safety and integrity of data, so it has triggered a great care from people. Nowadays the digital watermarking is developing quite quickly, and a lot of scholarly achievements have been obtained. Owing to the fact that the electronic file has its superiorities in efficient storage, retrieval, transmission and recondition,to protect the electronic documents are of great values. However, many documents contain non-text regions such as figures and pictures in addition to text. Thus, it's necessary to separate the document image and then embed watermarking in character area and image area.This article takes the image segmentation on document image watermarking as main research content. The main work is as follows:(1) The concept of digital watermarking has been introduced, and the basic frame of digital watermarking has been explained from the five fields of generating, embedding, extracting, detecting and attacking. The classifications, characteristics, typical algorithms and utilized fields of digital watermarking also have been introduced. Then the basic theory of document image segmentation and the typical segmentation algorithms have been introduced, the superiority as well as their limitation have also been discussed.(2) Analysis the character of document image. In space domain, analysis the gray histogram, mean, deviation, energy and entropy of the two different area; In frequency domain, analysis the mean, standard deviation, mean of the mean and standard deviation of the standard deviation of the two different area. Then a method of watermarking based on segmentation is proposed.(3) Propose an algorithm of page document image segmentation. Halftoning is first used to transform the gray-level image into binary mode. It can enhance the gray value of image area and weaken the gray value of character area. Use multiscale method to create character breviary image, to reduce computation complexity and character area information. And then a mathematical morphology method is used to make the image area is more integrity. Finally segment the character area and image area.(4) A semi-fragile DCT domain watermarking algorithm is proposed. The watermarking information is generated from the low-frequency domain, then embed into different position of middle-frequency domain based on different area of the document image. The experiment shows that, the segment method can effectively divide character area and image area from the document image. The PSNR of the watermarking image is improved in a certain extent. It can accept JPEG compression. The difficulty of detect watermarking is increased, and it also can accurately detect the tamper location.There are many incomplete problems of this algorithm. The according work will be done in the future.