Only graduate students taking 6.866 are required to turn in a final project (there is no final project for 6.801). This project for 6.866 is open-ended. Our goal with this project is for you to have the chance to apply what you learned in this course to solve a machine vision problem that interests you and expands on topics we’ve covered.

Suggested topics:

  1. “Time-to-contact” (TTC) and focus of expansion (FOE)
  2. Optical flow (either full variable flow or patchwise constant)
  3. Vehicle headlight tracking to obtain vehicle trajectories
  4. Line (not edge) finding plus vanishing points for rectification of head-mounted camera video
  5. Other machine vision algorithm (image/video input and on-screen output in graphic, text, or image format)

Suggested platforms:

  1. OpenCV (C++, Python) in Windows, Android, or another platform
  2. AndroidStudio with the Android SDK (The application itself will be programmed in Java and cross-compiled to run on an Android smartphone.)
  3. MontiVision in Windows
  4. Other platform able to support video input and output
  5. Python/C++

Grading Rubric

For 6.866, this final project comprises 33% of the final grade for this course. This 33% is allocated according to the rubric below.

Deliverables Percentages
Project Proposal 5%
Final Presentation 25%
Final Write-Up 70%

Project Proposal

Please keep this to 1 page, and all you will need to submit are a couple of paragraphs outlining your proposed work, any resources you will need/plan on using, and the specific deliverables you plan to accomplish. While we admire ambition, please make sure you keep your proposed work scoped enough that you will be able to finish it by the last day of class. The staff will follow up with feedback on your proposals.

Final Presentation

Final presentations will be held on the last day of class during our normal class time. Each final presentation will be a strict 5 minutes (you will have a 1 minute warning), with 1 minute of Q&A after. These will be conducted virtually over Zoom using our normal Zoom meeting link. Please have your slides uploaded before the end of class so we can reference them for grading, and please go through the following checklist to make sure we do not experience any delays as the timing will be a little tight:

  1. You will be presenting your slides, so please check your screen sharing capabilities prior to the presentation session to make sure they are functioning as expected.
  2. Have your slide deck (Google Slides, PowerPoint, etc.) pulled up and ready to present when it is your turn.
  3. Practice your presentation, and please keep it to a strict 5 minutes.

If for any reason you lose internet connection during the presentation, please do not worry. In the event that you are disconnected during your presentations, we will make accommodations and likely move you to the back of the presentation order so you can present again.

Note: If you are working on this project with another team member, please have each team member speak to approximately half of the talking points.

Your final presentation will be graded according to the following criteria:

  1. Coherence: How well did you communicate your project to the class?
  2. Organization: Is the presentation organized? Are the speaking points structured well?
  3. Connecting to Machine Vision: Did your presentation connect to what we have covered in Machine

Finally, out of respect for the presenters, please listen and engage for all the presentations.

Final Project Write-Up

A writeup no more than 4 pages (both single and double column are acceptable) is due after the project presentations. We encourage you to use LaTeX conference formats (such as CVPR), though this is not a strict requirement. We cannot give extensions to this deadline, so please plan accordingly.

Please structure the writeup like a conference paper, with the following sections included:

  1. Abstract
  2. Introduction
  3. Related Work
  4. Methods
  5. Discussion
  6. Conclusion
  7. References (does not count towards 4 page limit)
  8. If you work on a team, please provide a brief paragraph detailing the specific contributions of each team member (does not count towards 4 page limit).

This is not a strict structure that must be followed, but please make sure your writeup loosely reflects these sections.

Your writeup will be graded according to the following criteria:

  1. Is your methodology sound? Are your design choices and considerations clear?
  2. Is your methodology grounded in / related to concepts covered in machine vision?
  3. How well do you interpret your results? Even if your results are not the results you want or expect, please
    discuss them. We are not expecting you to break state-of-the-art with this project.
  4. How well does your related work put your project into context?
  5. How well do you frame your project and motivate it in your introduction? Try to think of specific domains
    and use cases for your project in this section.
  6. Does the writeup have a natural flow? Do the sections connect together? Is grammar generally correct?
  7. Do figures used enhance the quality of your writeup, and clearly communicate your project?

Please reach out to us if you have any questions or concerns about structuring your paper, and we would be happy to discuss more.