Research On Text Watermarking Algorithm Based On Hadamard Transform

Text documents have become one of the main forms of people's communication at present.And they can be spread on various storage devices in electronic form with the advancement of digital technology.But paper documents still occupy an important position in people's lives and work.The text watermarking technology uses human visual characteristics to embed hidden information as watermarks on the text carrier.When necessary,the embedded watermark can be accurately extracted by the corresponding extraction algorithm.The technology has extensive applications in copyright protection and traceability of text documents.At present,with the rapid development of input and output devices such as printers and scanners,as well as the popularization of mobile devices such as mobile phones,text documents will be subject to various attacks during the transmission of various devices,and then the accuracy of watermark extraction will be reduced.In terms of text watermarking,some methods have been proposed and corresponding results have been shown at home and abroad,but the visual effects are not ideal.Therefore,research on text watermarking that can resist various attacks is a promising problem.In view of the above situation,a text watermarking algorithm based on Hadamard transform is proposed,and a text watermarking system that can resist photographing is designed and implemented as well.The main contributions are as follows:(1)Aiming at sovling the problems existing in the previous text watermarking algorithm based on Hadamard transform,an improved text algorithm with better visual effect is proposed.First,a character division method is designed,which can be applied to a variety of texts,including different fonts,font sizes,word spacing,line spacing,and text lengths.The key issue is to count the length between different pixels to determine the step size for dividing characters.Then,adaptive embedding of characters is proposed.The first step is to calculate the maximum modification amount of each character.One can calculate the density of each character first,and then use the maximum modification amount of a character obtained by experiment as a reference,and finally calculate them in proportion.The second step is to control the distortion by making the actual modification amount of each character less than or equal to the maximum modification amount based on the different characteristics of different Chinese characters.Based on the above considerations,an optimization problem isproposed.That is,setting a variable to represent the real quantitative modification amount,and under some constraints,the maximum value of the ratio of the actual modification amount of the character to the maximum modification amount of the character is minimized.(2)Design and implement a text watermarking system that can resist attacks from mobile phones.The system is divided into two parts,namely the embedding part and the extracting part.The embedding part is composed of watermark information coding and watermark embedding.The work flow of the extraction part is to select the mobile phone picture,perform binarization,filtering and DPI matching,and then perform watermark extraction on the processed image,and the extracted watermark should be corrected for errors,and finally display the correct watermark sequence.In summary,the improved algorithm in this paper has better visual effect than the original algorithm,and can resist the attack from printers and scanners.In addition,the text watermarking system proposed in this paper can resist attacks from mobile phones and has good application prospects.