Course pages 2018–19

Advanced Graphics and Image Processing

• Two practical exercises and demonstrations, worth 30% of the marks.
• Mini-project, worth 70% of the marks.

Practical exercises

The details on practical exercises can be found in the Moodle course.

Mini-project

Students will write a short paper, maximum 4 pages, excluding the references, formatted in 2-columns. The short paper must be on one of the topics covered in the lectures (excluding the first lecture on "Introduction to image processing"). The paper can, but does not need to include supplementary materials: video, additional images, or a web page with additional results.

The selected research topic should involve re-implementing one of the published techniques and proposing improvements/modifications of it. It could also involve adapting an existing technique to a new application.

Project proposal

The students will submit to Moodle the title and 1-2 paragraph-long proposal of the research. The proposal should include 1-2 references to relevant papers.

Structure of the short paper

Most papers follow the structure with the sections: "Introduction", "Related work", "Method", "Evaluation" and "Conclusions/Discussion".

• The "Introduction" section should briefly describe the proposed technique/modification/application, but above all, it should motive the work. Why is this work worth doing and the paper worth reading?
• The "Related work" section should review 2-4 relevant papers, compare and contrast the proposed work to what has been published before.
• The "Method" section should clearly explain the technique, focusing on why certain decisions were taken, rather than just reporting the work that has been done. It should explain the method with the help of equations and illustrations, as appropriate.
• The "Evaluation" section should compare the method to the state-of-the-art, provide an evidence for improvement (or lack of it).
• The "Conclusions" sections should draw some insights from the work done, suggest future directions.

Mini-project time line

• 11/3/2019 - Project title proposal submitted to Moodle by 16:00
• 18/3/2019 - Topic approved
• 24/4/2019 - The short paper submitted to Moodle by 16:00

How to write an academic paper

The following references could provide useful guidance on how to write a good quality short paper:

• Write Good Papers and Notes on writing by Frédo Durand
• Modern myths – Shortcomings in scientific writing by Jean-luc Doumont

Assessment criteria

• 80-100% - Original idea and almost perfect execution; the paper could be accepted to a conference with minor modifications
• 70-79% - The work reproduces existing ideas/results, good execution and evidence of understanding.
• 60-69% - Like above but with occasional mistakes and faults or lacking clarity of presentation in some places.
• 50-59% - Somewhat incoherent, with important omissions, or irrelevant material. Some serious flaws in understanding.
• 40-49% - Work is poor and unstructured, with some parts missing. Little evidence of understanding, many inconsistencies and flaws.
• 30-39% - All aspects have been handled badly. Assigned work is substantially absent, incomprehensible or wrong. Little evidence of understanding, but at least mentions most of the relevant ideas.
• 20-29% - Almost total failure to engage with task. Assigned work is substantially absent, incomprehensible or wrong. Some reference to relevant ideas, but many are entirely missing.
• 0-19% - No evidence of any understanding. Marks may be given for presentation.