Robust watermarking techniques for authentication of printed copies through digital camera capture

Within the digital processing field, the aim of this thesis is to validate a robust image watermarking method and, afterwards, the authentication of these watermarks when images are printed and captured by a digital camera. This aim is focused to a specific business application: the sale of soccer sticker collections in both virtual and printed versions: printed images with the watermark will be captured by a digital camera of a smart-phone and later processed to detect the watermark. Thus, the purpose of the whole system is to analyze the robustness for uniquely identifying printed copies of those stickers for fan collectors. The work is organized into four chapters. The first chapter shows some concepts about digital watermarking, such as a definition, main properties, a theoretical model and a classification of the types of watermarks. Afterwards, it is presented the selected watermarking technique, how it works and also the advantages and disadvantages of this selection. The second chapter details the algorithms for insertion and detection of invisible watermarks that have been developed, and the implemented routines in a programming language according to these algorithms. These program routines have been developed in MATLAB programming language. The third chapter presents some experimental results and also conclusions extracted from them. The different tests detailed in this chapter may allow a more accurate use of the developed functions and a better knowledge of their behavior under specific possible circumstances. Finally, the last chapter concludes and summarizes the work done in this thesis and gives some suggestions for future work in following studies