Group Project

Rules of the game

  1. Work in teams of 3-5 people; 3-4 is recommended.
  2. Prepare an A1 poster on a chosen topic, relevant to the module. The poster will be given a collective mark (to all team members) worth 25% of the module.
  3. Present the poster to class during the sessions in Week 11-12. Each team member has to be active in the presentation, and will be given an individual mark worth 25% of the module. One possible approach is that each group member present one key point made in the poster, and relate it to the rest.
  4. At most two teams will be allowed to do the same (or strongly overlapping) topics. First come, first served applies.

Deadlines

7 February 10am
Team up and submit list of group members. I recommend that you discuss when you will do the work before you agree to work together. In particular, some may want to work on the poster during the Easter breaks while others go away on holiday.
21 February 10am
Suggest topic. A detailed project description is not necessary at this stage, but please include all the details that you have, especially all references.
6 March (office hour)
Agree topic with module convener. If an agreement has not been made before 6 March, you should meet in person to discuss it during my office hour. (This deadline leaves one week to answer any questions before the Easter break.)

What a good poster looks like

  1. A good poster can be read at a distance (at least 3m). Don't use too small fonts.
  2. A spectator should get useful information at a quick glance.
  3. Use as many visual aids as possible.
  4. Don't say too much. If you try to make one point, you can hope that it is remembered. If you try to make six points, the reader is lost and remembers nothing.
  5. Coherence and consistency. The entire poster should build up under the same message. Again building up under one big point.
  6. Quality is better than quantity. The assessment will focus on your ability to prioritise material, picking the most useful information, to relate information from different source, and to relate this information to real-world applications.

One sub-point (each building under the main point) per team member makes it easy to split the task for the verbal presentation.

Sample topics

This list will be revised until the deadline for forming the groups. Most importantly, I will add possible papers for review and possible systems for implementation. You are welcome to suggest your own topics.

  1. Consider steganography by mimicry using grammars, as described in Wayner's book and implemented by spammimic.com. Is it a secure? Is it practical? You can (for instance) either (1) evaluate the idea, possibly by implementing and testing it yourselves, or (2) evaluate the spammimic.com web site.
  2. Implement, test, and evaluate steganography in synthetic English text based on frequency analysis, as described in Wayner's book.
  3. Review a relevant paper which has not been properly covered by the lectures. In many cases, it is a good idea to implement and test proposed algorithms.
    • Backes and Cachin 2006 (LNCS vol. 3378)
    • Hany Farid, Information Hiding 2002 (see LNCS)
  4. Implement, test, and evaluate at least one stegosystem/steganalysis technique which have not been covered in the lectures.
    • F5 and Matrix embedding
    • LSB with Matrix embedding
  5. Give an introductory overview to one (1) of the following subjects and its relations to steganography, adding some interesting detail beyond the coverage from my lectures.
    • Statistics: estimation and hypothesis testing.
    • Information Theory and Data Compression
    • Coding Theory
  6. Make a literature survey on applications of steganography and steganalysis in `terrorist' organisations and intelligence services. Assess the significance of the field in terms of national security. You should make an extensive literature search, but avoid technical detail.
  7. What is randomness? Such an apparently simple question can be quite controversial and philosophically challenging. The project should browse the literature for a range of opinions about and attempts to define randomness. Subsequently, these findings should be interpreted and summarised with a view to applications in steganography.
  8. In principle LSB and other spatial stego-techniches by modification apply also to palette images (e.g. GIF), but there are certain problems. Explain what palette images are, and discuss how steganography by modification can be performed. It is a good idea to base the disussion on own implementations and tests.

Assessment Criteria

Poster

less than 40
Clearly failed to reach the learning objectives.
40-50
Some descriptive understanding is demonstrated, but the poster does not sufficiently cover the agreed material.
50-60
The poster shows a declarative and/or procedural understanding of the material (paper/method) studied.
60-70
In addition to the above, the poster has a clear and coherent focus, where the different pieces of information are clearly prioritised
70-85
In addition to the above, the paper views the material in a wider context. This could mean, for instance, comparing the presented method/work with other sources (such as lectured material); assessing the presented work in view of real-life applications.
85-100
Beyond any expectation

Presentation (individual)

less than 50
Unclear contribution or understanding.
50-60
Clear presentation of own contribution (part of poster). Descriptive understanding is demonstrated.
60-70
Well-structured presentation with a clear focus, meeting the requirement above, but short of the 70+ band.
70-85
The presented part of the poster is well presented and clearly related to the poster as a whole.
85-100
Beyond any expectation