Design this project yourself. Pick something about which you are interested in learning more. Once you have an overall topic, pick a specific task or application. Plan out what resources you will need and how long you think it will take. The overall scope of the project should be similar to the other projects. Keep it simple.
You may work in groups of up to four people. Groups larger than two will have higher expectations.
The bulk of your grade will consist of your report, your presentation, and well-commented and organized code. The quality of your results and the scope of the project will determine the final 15%.
The final project is due on Wednesday April 26, 2023. You cannot use time travel days on the final project as course grades are due on Monday May 1st, and we need time to review and grade the projects.
Note the check-in date.
-
Project Preparation As part of project 5, you should have completed your final project proposal.
-
Complete your project Checkin 04/14/2023: one person in your group should send the professor an email with a 1 paragraph project description, the names of your group members, and a progress update.
-
Write a report Your report should be a pdf in IEEE 2-column conference paper format. Templates for Word and LatexLinks to an external site. are available from IEEE. Your report should contain the following sections:
- Abstract, 200 words maximum
- Introduction section: describe the task and the main concepts at a high level.
- Related work: find at least 3 peer-reviewed papers that are related to your project. Give a brief description of each and how it relates to your project. CVPR, ICCV, ECCV, BMVC, and WACV are all computer vision conferences that are good sources for papers. Most are available as pdfs through the Computer Vision Foundation [CVF]. Many papers are available via arXiv, but please find the peer-reviewed venue where they were actually published.
- Methods: describe your approach to the problem and the specifics of your solution at a level of detail that someone knowledgeable in the art could reproduce it. If you are following someone else's work, be sure to cite the appropriate paper.
- Experiments and results: describe what experiments you undertook to demonstrate your system worked and what the results were. If there is comparable prior work, compare your results to the prior work. Graphs or tables are good for presenting results.
- Discussion and summary: discuss your results and any key ideas or findings from your work.
- References/Bibliography Your report should no more than 8 pages, including the references and any diagrams or pictures. It can be less.
-
Generate a video presentation for your project Record a presentation that is no more than 15 minutes long to present your project and the results.
- The presentations should be no more than 15-minutes long.
- Each person in the group needs to be part of the presentation (e.g. presenting at least one slide).
- Please prepare a set of slides. For a 15-minute presentation, you should have no more than 8-9 slides, including the title slide.
- The presentations should include an introduction, a brief description of related/prior work, your own solution, and the results. Include a demo in your presentation if possible.
There are no extensions for this project. Do more if you have the time.
Submit your pdf report, code, and a readme file to Gradescope. When you submit to Gradescope, be sure to add all of your group members to your submission. Please do not submit videos or data sets to Gradescope. It does not handle large files. Instead, put them on Google Drive or YouTube and provide a URL. You do not need to submit your data sets, but you may want to have them on some kind of cloud storage for later use.
In your README file, please include the following in addition to any other relevant information.
- Your name and the names of all of your group members.
- Your project description (same one you used for the email check-ins)
- The URL for your presentation
- The URL(s) for any other demo videos you might have.