AL ZAID, A. Digital Image Watermarking [online]. Brno: Vysoké učení technické v Brně. Fakulta elektrotechniky a komunikačních technologií. 2009.

Posudek vedoucího

Číka, Petr

The work activity of Abdullah Saleh Al Zaid was very good during the whole year. He proposed 2 methods for still image watermarking in the frequency domain. These methods were compared with each other using the Checkmark benchmark tool. All results are listed in tables. The thesis is very good in formal aspect. The figures describe the problematic and suitably supplement the text. I recommend this work to be evaluated with mark A/95 points.

Posudek oponenta

Říha, Kamil

The reviewed work is of relatively good-quality, especially almost all the graphical items are very precious. Even the volume of the work is more than sufficient. The overall concept of the text structure can be found as uselessly structured (e.g. subchapters in chapter 3.2 are numbered as chapters although they contain 1-4 sentences), there can be found also chapters with almost identical texts (chapters 7.2., 7.4.). The language of the work is good, except for some sporadic, hardly intelligible formulations and language discrepancies. By the examination of the thesis content it is possible to find some insufficiencies or inaccuracies. For example, Fig. 18 bottom left should represent a matrix of DCT coefficients of one 8x8 block and not the source image (see Question 1). Reactions of the two digital image watermarking techniques to a number of different attacks are evaluated and quantified by tables for each method separately although a graphical or text (as stipulated in the assignment) comparative analysis would be suitable. In conclusion I can observe that the work fulfils the requirements for bachelor thesis with smaller reservations and I suggest evaluating it by mark 84 in a grade scale from 0 (bad) to 100 (excellent).

eVSKP id 28327