Research On Database Watermarking Algorithm Based On Differential Expansion And LBP Feature Extraction

The advent of the era of big data provides a broader stage for information transmission and information sharing,as well as a chance of a hundred years.At the same time,the security of data in the process of transmission also facing an unprecedented severe test.Many data are related to the national economy and the people’s livelihood,scientific research patents,military defense,privacy security and other important fields.Once these data are stolen or destroyed by lawbreakers,it will bring great losses to the country and the people.Therefore,it is particularly urgent and important to protect the security of the database.The key to solving the security problems of data(i.e.copyright protection,information preservation,etc.)is to choose the transmission mode and carrier of data,Database watermarking technology is just based on the way and carrier of data transmission,which can solve the problem of data being tampered with at will and important information being leaked,and can effectively promote the safe interaction of data and information transmission.Although database watermarking technology is used as the "firewall" of database,there are still some problems to be solved,such as weak resistance to attack,small capacity,and difficult to embed.In this case,this dissertation proposes a database watermarking algorithm based on differential expansion and artificial firefly swarm optimization technology,and adatabase zero watermarking algorithm based on LBP features and hash function,which improves the ability of watermark to resist attack.These two algorithms expand the embedding capacity,reduce the embedding time and reduce the embedding difficulty under the premise of ensuring the concealment.The database watermarking algorithm based on differential expansion andartificial firefly swarm optimization technology proposed in this dissertation first identifies the database by a hash function,then divides the database into non-overlapping subgroups according to the length of the watermark,and then passes the artificial firefly to the subgroups.The group optimization algorithm selects the best attribute column,and finally embeds a "0" or "1" watermark into the database dynamically through the differential expansion method.The majority selection method is used for extraction.This algorithm can embed the watermark capacity up to about 37.8%of the total number of tuples under the premise of different database tuple sizes to ensure the concealment of the watermark,saving the embedding time and improving the robustness of the watermark.The database zero watermark algorithm based on the LBP(Local Binary Pattern)feature and hash function,first we should select the feature attribute column through the Pearson correlation coefficient,then divide the selected attribute value into the part to be embedded(least significant bit)and the part to be extracted,and apply the data to the extracted part after the local binary pattern(Local Binary Pattern),the perceptual hash is used to generate the 0 and 1 sequences,and then the XOR with the base watermark generates the zero watermark.Finally,the generated zero watermark feature code is embedded in the least significant bit of the selected attribute column using a database watermarking algorithm based on a hash function.In this process,the hash function is used twice to improve the concealment of the algorithm.In the watermark extraction,the majority voting method and the Hamming coding method are introduced to control the watermark error.Experimental result shows the algorithm guarantees the availability of the database after embedding the watermark,and after being attacked by various methods,the robustness of the database watermark has improved to a certain extent.The correct extraction rate of the watermark in the zero watermark mode is about 90%.The two algorithms propose in this paper provide a new idea for database copyright protection.Based on the differential expansion and artificial firefly swarm optimization technology,the database watermarking algorithm can recover the original data while extracting the watermark.Under the premise of ensuring concealment,the embedded capacity is Improved;database zero watermarking algorithm based on LBP feature and hash function,using local binary mode to generate zero watermark,and then embedding the generated zero watermark in the database to play the role of copyright identification,the combination of the two can achieve double layers The role of protection.