|
AIM
|
- |
The Center for Biological & Computational Learning is offering a novel UROP position in the field of computer vision. The goal is to implement a fast version of a computer vision algorithm using ast parallel operations on a modern GPU. In the end, we should like to produce an approximate real-time system. It is likely that a completed successful work would result in a publication.
|
|
TECHNIQUE
|
- |
The student will first read and understand the computer vision model developed here at CBCL. Secondly, the student will investigate hardware-level programming techniques for our GPU. Finally, the model will be re-implemented to run as fast as possible. The algorithm is well defined from a high level, but there is plenty of room for creativity in the utilization of the specialized GPU hardware.
|
|
PREREQUISITES
|
- |
(1) MIT undergraduate student.
(2) Capable of programming to the hardware for a modern GPU.
(3) A minimum of 15 hours per week of lab time is expected.
|
|
|
|
REIMBURSEMENT
|
- |
Hourly rate or credit as requested.
|
|
START TIME
|
- |
ASAP |
|
DURATION
|
- |
One semester |
|
CONTACT
|
- |
Stan Bileschi or Lior Wolf
bileschi@mit.edu, liorwolf@mit.edu
Center for Biological & Computational Learning
http://cbcl.mit.edu/jobs/index-jobs.html
|