EECS 755, Class Project, F13 Goals: Deeper understanding and hands-on experience with image reconstruction. To help improve project skills: project proposing, independent learning, project execution, and project presentation (both oral and written). (In an advanced special topics course like this, it is more natural to use the subject material to solve real problems than to try to synthesize exams.) The nature of projects: A typical project would involve learning about and experimenting with some image reconstruction technique. For example, a project might involve studying and investigating some technique introduced in class or in the book. Numerous "open problems" are mentioned in the book as possible starting points for projects. The usual goals of image reconstruction research are to develop methods that lead to improved image quality, or to develop algorithms that provide the same (or similar) final image quality but reduce computation time. An example project would be to compare two (or more) algorithms for minimizing an image reconstruction cost function in an application area of interest to you, or to compare the image quality that results from different choices of regularizers or other algorithm parameters. Most projects will probably include numerical experimentation using Matlab, perhaps using the tools in the Image Reconstruction Toolbox (IRT). However, purely theoretical projects (convergence proofs, analysis of algorithm properties, etc.) are also welcome. Projects may closely follow existing reconstruction methods from the literature. You need not invent something new. There is often plenty of room for experimentation, analysis and design, even with existing methods. On the other hand, many past project have led to conference papers and journal publications. "Deliverables" (Your project grade will be based on these three items.) 1. (0%) Project proposal: Recommended deadline: Nov. 1 I highly recommend that you first discuss with me in person your initial plans for the project in October so that I can give you initial feedback and suggestions of references. Based on this feedback, I recommend that you prepare and submit a 1-2 page description of your plans for the project, including references. Upload a pdf file to ctools (the system will email me that you uploaded it). I will give feedback based on your proposal. 2a. (10%) Oral presentation: At the end of the semester, each student will give a 10 minute oral summary of their project to the class, followed by a few minutes of questions from the class. The format of these presentations will be just like at an IEEE conference. The presentation should be targeted at students in the class, i.e., you should assume the audience has knowledge of general advanced image reconstruction methods but may not know the details of your application / algorithms. These presentations will be at the final exam time: Dec 17 1:30PM-3:30PM, in EECS 3433. 2b. A pdf of your presentation, uploaded to ctools by 1:30PM Dec. 17. 3. (20%) Written report: Upload pdf to ctools by 5:00 PM on Dec. 17, preferably, but I will accept it up until 5PM on Dec. 18 because that was the time posted originally. The format of this report should be a 4-5 page IEEE conference paper (longer is OK but not required). The audience again would be attendees at a conference that is suitable for the topic, e.g., ISBI. No Matlab code need be included in this report, but if you have developed an algorithm that would be a useful addition to the IRT, then please email it to me separately. In both your oral presentation and your report, if you cut-and-paste in figures / images that you found on the web, then you must cite the source. Ideally, the report you write for this class will not be a mere exercise, but rather will be a draft of something that you will in fact submit later (perhaps with additional investigation and results) to a conference. Scoring factors include: quantity, quality, degree of difficulty of the work. The analytical or experimental results. Connectedness to the course. The quality of your critical judgments and analyses. The depth of understanding displayed. Innovativeness. Grading will take into consideration the course background of the student.