Overview
The report should be written for an audience of EECS 556 students. Assume that the students understand the class material, but have not seen the new material that you have investigated. They should be able to learn your topic from your report. The description should be sufficiently detailed that they could easily repeat what you have done. No code should be included in the report; instead the algorithms should be described mathematically or with pseudo-code, like in a journal paper. Do discuss any unusual commands used in your project, especially any tools you got from the web. A typical report might be 10-12 pages plus figures, tables, and references, but there is no page limit. Concise and thoughtful summaries of your results are more desirable than a "laundry list" of example images.
Use your existing overleaf document as the starting point. I recommend you pretty much leave everything in there and *add* to it the Abstract and the Results. Change phrases like "We will implement..." to "We implemented..." so that it is a report, not a proposal!
Typical format of report
Title, Authors, Class, Date
(The title need not match the paper you reproduced.)
Use full names, not initials - own it!
Honor code
(Be sure to follow it when using images and figures)
Abstract
1 paragraph overview of project
(I will need an email copy/draft of this paragraph
*before* the project presentations.)
Introduction
Describe problem and why it is interesting and important.
Summarize some of the types of methods that have been applied
to this problem previously (if applicable).
Cite the key paper that you used for your project.
Give a qualitative description of the methodologies you investigated.
Typically there will be very few if any equations in this introduction.
Quantitative performance prediction
Restate your quantitative performance prediction
from your proposal.
Methods and/or Theory
Describe the algorithm(s) that you developed/implemented/investigated.
Use mathematical descriptions as necessary to describe the methods.
For "simple" steps you can just use words like
"we found the coefficients by ordinary least squares fitting"
or "we used a 2D FFT to determine the spectrum."
But do not just say "we found the coefficients" because that is too vague.
For more involved steps you need to give more specific detail
about how you designed/implemented the method(s).
For the non-obvious steps specify whether you implemented the method yourself
or used software from the web etc.
Use active voice: "we implemented ... we analyzed ..."
because then it is clear who did it.
If you write in passive voice
(e.g., "the method was implemented using a loop")
then it is unclear whether you did it or whether someone else did.
Experimental results
Describe how you evaluated (or analyzed) the performance of the method(s).
Specify what datasets used
(where did they originate, citing references / urls if needed)
and what metrics you used to assess or compare methods.
Use active voice: "we compared ... we evaluated ..."
The default font sizes in Matlab are usually *too small*
for presentations and reports.
Make sure you make the fonts big enough to be easily readable!
Please make the images big in the report
so that it is easy to see the details.
I will not print the reports so no paper will be wasted by making big images!
This is an image processing class
so I certainly expect to see some representative
image results in there.
For some applications,
image axes with units are needed
and/or colormaps are needed to show quantitative image values
(like depth maps).
Use your judgement.
Make sure that any images that you import into the report are
not corrupted by JPEG blocky artifacts.
Your restoration method will not look very impressive
if all I can see is 8x8 blocks...
Conclusions and future work
Summarize the main conclusions of your project.
In particular,
describe how the results compare
to your original quantitative performance prediction.
Recall that your prediction was supposed to be based
on results in the key paper;
discuss the paper's conclusions in light of your results.
Try to briefly explain
why your performance was better or not than the prediction.
Your score here will depend on your discussion
of the prediction and results,
not on whether the results matched the prediction.
List some areas for further improvement or future research based on what you learned and/or what challenges you encountered.
Bibliography
See technical writing guidelines.
Code
Most groups (hopefully all groups!)
are using a private github repo
for the code in your project.
As you finish your report,
please add me to access to that repo
and include the url
to that github repo
at the end of the report.
If you do not have a github repo
then put the code in an organized google drive folder
and share the folder with me
and also put the url of that folder
at the end of the report.
Whether you use github or google drive,
be sure to include a README.md (github)
or README.txt (google)
file that explains the organization of the code.
Here is a
template for good README organization
recommended by NeurIPS.
Group effort table
The last page of the project report
must include a table reporting group member effort,
by incorporating
this .tex file
into your overleaf doc.
The result will be a nice table that is
similar to this one.
Be sure to read and follow the instructions!