Embedding watermarks in multiple tables in numeric dataset by bit pattern generation sub-watermarks

Abstract

A novel and technical approach for watermarking tables of a numeric dataset is proposed. Watermark insertion and detection is done by K Map Minimizer. The crucial aspect of the proposed technique is the permuted bit pattern generation. The bit pattern is generated by first selecting unique rows and columns and solving the four attributes at the selected position through the K Map Minimizer to constitute a pattern. To make the bit pattern more secure permutation is applied before, where the size of the bit pattern is adjusted as per requirement. This unique permuted bit pattern is called the watermarked permuted bit pattern. Each bit of watermarked permuted bit pattern is added /subtracted while sub watermarking the selected attribute. Four secret keys and one permutation key are used which adds to the integrity of the technique. Secret keys are always used in pair so knowledge of a single key to an attacker, is of no use rather access to k-1 keys, where k is the number of keys, is also of not much worth as to generate a watermarked permuted bit pattern all four keys are requirement. Watermark detection routine is blind as it neither requires the original dataset nor the watermarked one. Watermarks are equally distributed in a very large dataset as six digits number is selected through K Map Minimizer, so the watermark can be detected even in a small subset of a watermarked relation as long as the sample contains some of the marks. Proposed technique is robust against various malicious attacks and updates to the data. Dealing with multiple tables adds to the practicality of the data.