In order for this to work, Pirsiavash and Ramanan have created a library of algorithm training examples of videos depicting a particular action, and specified the number of subactions that the algorithm should seek. The rules relating to subactions are the key to the algorithm’s efficiency, according to MIT:
As a video plays, the algorithm constructs a set of hypotheses about which subactions are being depicted where, and it ranks them according to probability. It can’t limit itself to a single hypothesis, as each new frame could require it to revise its probabilities. But it can eliminate hypotheses that don’t conform to its grammatical rules, which dramatically limits the number of possibilities it has to canvass.
The researchers tested their algorithm on eight different types of athletic endeavors with videos from YouTube and found that it identified new instances of the same activities more accurately than its predecessors. David Forsyth, a professor of computer science at the University of Illinois at Urbana-Champaign not involved in the research said in the release that there was been a strong sense in the community that this sort of algorithm should work, and that to date, this is "certainly the best of this sort that I’ve seen."
With an eye toward its potential applications, Pirsiavash suggests the algorithm could be used for medical purposes, including proper execution of physical therapy or determining whether elderly patients remembered to take medication. Other potential applications of this algorithm could include security and surveillance applications, but where else could you see this being used? Let us know in the comments below.
View the MIT news release.
Also check out:
MIT researchers develop algorithm for better robotic navigation and scene understanding
Developers look to open sources for machine vision and image processing algorithms
Software aims to characterize algorithm performance
Share your vision-related news by contacting James Carroll, Senior Web Editor, Vision Systems Design
To receive news like this in your inbox, click here.
Join our LinkedIn group | Like us on Facebook | Follow us on Twitter | Check us out on Google +
Page 1 | Page 2
