CRA-W DMP Summer 2006

Erika Chin

Seattle, WA

My Homepage My Mentor My Journal My Final Project

Greetings! I am Erika Chin. This coming school year I will be a fourth year undergraduate student in the School of Engineering and Applied Sciences at University of Virginia. I am class of 2007 and I am currently working towards a BS in computer science and a minor in engineering business.

This summer I was chosen to be a participant in the Distributed Mentor Project (DMP) sponsored by the CRA-W. This is a program that takes undergraduate women in computer science and engineering and matches them with a faculty mentor for the summer. Many universities all across the country participate in the program. It is a wonderful way for an undergraduate student to learn more about the research process, visit another university, and experience life as a graduate student.

I was paired up with Maya Gupta, a professor at University of Washington (Seattle) in the Department of Electrical Engineering. She does research in the areas of estimation and classification, image processing, information theory, and the science of color vision. For more information on this, please see the Information Design Lab webpage.

This summer I will be working on inverse color management. Inverse characterization transforms compensate for the unique color characteristics of a device and they allow for the input into a printer to match the output of the printer. I will specifically be working on ways to take a few sample pairs (training data) and generate a lookup table that maps CIELab to RGB. This way, when a person knows what color they want the printer to reproduce, they simply need to take the CIELab (which is a device-independent color space) and use the tables to find the RGB values (device-dependent) that must be sent to the printer. Within the inverse color management process, I will experiment with different adaptive, enclosing neighborhoods and use different regression estimation techniques on those neighborhoods to derive RBG values. The produced look up tables can then be used to generate an ICC profile for the printer.

Through this experience, I hope to: