The type of the final project may be one of the followings (but is not limited to): a survey
or review paper, an implementation of an existing algorithm, a performance characterization of some
existing techniques, an improvement of the existing techniques, and a new algorithm based on  
combining vision theories with a deep learning model, with a focus on incorporating theories into 
the deep model.

Students should consult with the instructor as early as possible on the topics of the 
final project.
The project grade of a student will be
primarily based on an oral presentation, a written report, and the source code if applicable. 
The report should have a
minimum length of 10 pages including an introduction, literature review, problem statement, 
details of
the proposed approach, performance evaluation of the proposed approach, and finally a conclusion. 
Each
student is expected to turn in a project report by the end of the semester. No late reports will be
accepted under normal circumstance. The source code should be well documented and should be
deposited in a designated area. 

Two criteria I use to evaluate
students' project proposal. First, it shall relate to what you
have learned in class. Specifically, it shall either apply the 
knowledge learned in class or expand it. 
Second, your need explicitly state the scope of
work, based on which I will evaluate your work.