Image authentication and recovery: Sudoku puzzle and MD5 hash algorithm based self-embedding fragile image watermarking method

Renklier A., Öztürk S.

MULTIMEDIA TOOLS AND APPLICATIONS, vol.0, pp.1-23, 2023 (SCI-Expanded, Scopus)

  • Publication Type: Article / Article
  • Volume: 0
  • Publication Date: 2023
  • Doi Number: 10.1007/s11042-023-15999-2
  • Journal Name: MULTIMEDIA TOOLS AND APPLICATIONS
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, FRANCIS, ABI/INFORM, Applied Science & Technology Source, Compendex, Computer & Applied Sciences, INSPEC, zbMATH
  • Page Numbers: pp.1-23
  • Kayseri University Affiliated: Yes

Abstract

In recent years, image forgery has become an important issue with the use of images in many important areas. Several self-embedding fragile watermarking methods have been studied for image authentication with the additional capability of image recovery. However, these methods have certain shortcomings in terms of coincidence problem, imperceptibility, and image recovery. In this paper, we introduce a new self-embedding watermarking method based on the Sudoku puzzle, in which recovery is achieved by successfully reconstructing up to 80% tampered images. In the proposed method, initially, the grayscale host image is divided into twenty-five blocks based on the 5×5" role="presentation" >5×5 Sudoku puzzle. These blocks are grouped and divided into 5×5" role="presentation" >5×5 non-overlapping sub-blocks. The recovery information is obtained from the average pixel values of each sub-block. For each group, four copies of the sub-block recovery information are embedded into the second least significant bits (LSBs) of the other group blocks based on Sudoku. Then, for each 5×5" role="presentation" >5×5 sub-block authentication information is generated with MD5 hash algorithm using the block position, block pixel values, and a secret key. Finally, the 25-bit of the authentication information is embedded into the first LSBs of the sub-blocks. Experimental results show that the proposed method has good performance in different sized cropping, splicing, copy-move, salt-pepper, histogram equalization, sharpening, blurring, and text-addition attacks applied to different regions of the images. Also, the proposed method has obtained high quality recovered images since it efficiently minimizes the effect of coincidence problem.