Issues in integrating and implementing an open source toolkit and demonstration system for digital image watermarks

Abstract

Free/Libre and Open Source Software (FLOSS) has been an important factor in the software world and it has profoundly changed the way people deal with software products. However, it seems that little job appear to have been done to develop digital watermarking systems as FLOSS. Developing such a system requires multiple-disciplinary knowledge of digital image watermarking techniques, digital image processing, information hiding, computer science technology and cryptography. The demands of these diverse subjects may suggest that building such a system is a difficult task. This thesis explores the feasibility of setting up an Internet-based digital image watermarking system capable of embedding and extracting text messages as FLOSS. Several digital watermarking algorithms, such as Least Significant Bit, Patchwork, brightness modulation and DFT amplitude and phase modulations were explored. A choice was made to adopt a J2EE (Java 2 Platform Enterprise Edition) browser/server structure to build the embedding/extraction subsystems and to use Java, HTML/XML and Javascript languages to implement the system, as these are independent of specific computing platforms (eg, Windows or Linux). A trade-off analysis on watermark imperceptibility, robustness, reliability and capacity of algorithms was carried out and taken into consideration in shortening the processing time in the empirical system. The thesis work has developed an illustrative framework for future extension and research.