Title: A unified technique for robust digital watermarking of colour images using data mining and DCT

Authors: Samir B. Patel, Tejas B. Mehta, Shrikant N. Pradhan

Addresses: Computer Science and Engineering Department, Institute of Technology, Nirma University, Ahmedabad 382481, Gujarat, India. ' Computer Science and Engineering Department, Institute of Technology, Nirma University, Ahmedabad 382481, Gujarat, India. ' Computer Science and Engineering Department, Institute of Technology, Nirma University, Ahmedabad 382481, Gujarat, India

Abstract: Digital watermarking is an emerging copyright protection technology. The paper presents a new robust watermarking technique based on combining the power of transform domain technique; the discrete cosine transforms (DCT) and the data mining technique such as decision tree induction (ID3). The paper focuses on a technique through which the notion of decision tree can be applied on transformed vectors to build the decision tree. We train the image blocks for deriving the classification tree. The resulting decision tree provides decision-making rules to identify good quality image blocks for insertion of watermark. The implementation results have shown that the algorithm has an acceptable robustness against the JPEG compression and addition of noise.

Keywords: digital watermarking; discrete cosine transforms; DCT; Arnold transform; data mining; decision tree induction; ID3; security; colour images; copyright protection; digital watermarks; JPEG compression; noise.

DOI: 10.1504/IJITST.2011.039680

International Journal of Internet Technology and Secured Transactions, 2011 Vol.3 No.1, pp.81 - 96

Published online: 29 Nov 2014 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article