Research:
I use ideas and techniques from Bayesian statistics and machine learning to construct models of perception, both to investigate the structure of the noisy world around us, and to construct systems capable of perception and action. Recently, I've been investigating how we can recognize events in video using a combination of low level motion (optical flow of salient features) and external intelligence (object recognition) to discriminate between different activities.
Here's an example of the motion flows:
Here's the same flows with no bakcground:
Here are some flows labeled by the mixture model described in the AAAI Spring Symposium 2009 paper:
The work that these movies display is described in:
Messing, R. & Pal, C., 2009 "Behavior Recognition in Video with Extended Models of Feature Velocity Dynamics" AAAI Spring Symposium 2009. PDF
My advisors are Henry Kautz and Chris Pal. I'm working on using motion to classify activities of daily living (ADL) as part of an assisted cognition system.
Frequently Used Links:
My undergrad work has been published:
Messing, R. & Durgin, F. H., 2005 "Distance Perception and the Visual Horizon in Head-Mounted Displays" ACM Transactions on Applied Perception, 2(3), pp.234-250. External Link
Email: The first letter of my first name, followed by my last name, at cs.rochester.edu. This is the best way to reach me.
Office: CSB 631
Office phone: 585-275-0469