Undergraduate Projects

Mable Zhang

Identifying the Cuisine of a Plate of Food

StudentsMabel Zhang
Year: 2011
Where are they now: UPenn
Quote: "Working on a project that I designed by myself gave me a lot of motivation, because it was something I personally like to see happen. Different from real research work I've done in professors' labs, in this project, I got to pick the topic I want, collect data the way I like, and then have fun with it at my own pace without being pressured with deadlines. I felt I had complete ownership over the project, and with the guidance in the class, it was much easier to push myself to accomplish something in a limited time. Starting it from scratch to working taught me a lot about Linux, OpenCV, and memory management that I never had to worry about before. It is definitely a worthy investment with free help."
Keywords: food, classification, attributes

Food

two men greeting each other

Nasolabial Wrinkle Classification

StudentsDavid Lluncor
Year: 2010
Where are they now: Google
Quote: "CSE 190 is a class that I recommend for anyone who is serious about being involved in undergraduate research in the Computer Science department. It provides an invaluable chance for you to interact with students who are completing projects at the same pace you are, and at the same time receive feedback from professors in an informal and relaxed setting. This class keeps you on your toes and is a great launching pad for delving deeper into a particular topic in Computer Science. Highly recommended for people who have the slightest inclination to complete an independent project, work with friends, and really learn a topic such as Computer Vision better."
Keywords: dermatology, facial imaging

Graph

Dafna Bitton

Recognition of Handwritten Names

StudentsDafna Bitton
Year: 2009
Where are they now: Amazon
Quote: "This was a great opportunity to work on a project that was driven by myself and that I had an interest in. The class was good motivation to keep on track with the project and both the students and instructor provided valuable feedback to my weekly progress."
Keywords: boosting, ocr

Bright early grid

 

Thavidu Ranatunga

Rat’s Life Competition Entry

StudentsThavidu Ranatunga
Year: 2008
Where are they now: Microsoft
Quote: "This was my first time doing an individual research project and it definitely got me interested in doing more projects later. It was an invaluable start =) It was a great experience and I definitely learnt a lot, especially from sharing with what the other exciting projects were doing in different areas of CV. Highly recommend it!"
Keywords: robotics, mapping, sensors

Rats life

Shane Grant

A Modified Approach to Biologically Motivated Saliency Mapping

StudentsShane GrantKevin Heins
Year: 2008
Where are they now: USC
Quote: ""I learned things I probably never would known about or been taught because of the projects I worked on. It was a great experience to be able to start with set of simple ideas and see them grow to fruition, getting invaluable feedback from interaction with my peers and the instructor along the way." -Shane "
Keywords: visual saliency

Balloons

 

Nadav and Boris

ReSPEC

StudentsNadav Ben-HaimBoris Babenko
Year: 2006
Where are they now: Orpix, Inc. / UCSD graduate student
Quote: "This research project got our 'foot in the door' of computer vision research at UCSD. We went from having 1 image processing course worth of experience in the field to publishing a workshop paper which currently has 23 citations."
Keywords: content based image retrieval

Computer vision flow

Fred Birchmore

Detecting Soda Cans Against a Cluttered Background

StudentsFred Birchmore
Year: 2006
Where are they now: SPAWAR
Quote: "This project enabled me to make the transition from student to researcher and completely transformed the way I learn in general. As a researcher, it is important to get through the coding and straight to the experimental portion as quickly as possible. If you are 'digging in the wrong area', you want to find that out before investing too heavily in software development."
Keywords: object detection, keypoint matching

Test cans