Action, scene, and audio concept annotations on TRECVID2010 MED development videos. 

	Provided by Digital Video and Multimedia Lab at Columbia University. 
  
  Citation: 
  Yu-Gang Jiang, Xiaohong Zeng, Guangnan Ye, Subh Bhattacharya, Dan Ellis, 
  Mubarak Shah, Shih-Fu Chang. Columbia-UCF TRECVID2010 Multimedia Event Detection: 
  Combining Multiple Modalities, Contextual Concepts, and Temporal Matching. 
  In NIST TRECVID Workshop, Gaithersburg, MD, November 2010.
  
	Please contact Yu-Gang Jiang (yjiang@ee.columbia.edu) for questions and bug reports.
	1/23/2011.

-----------------------------------------------------------------------------
We annotated 565 videos in total. Labels are provided at 10-sec clip level.

Label file (med10conceptLabel.txt) format:
	
	+ each row is a 10-sec clip; columns are:
		- MED 2010 video ID
		- start time of the 10-sec clip
		(followed by binary labels of 6 human actions)
		- Person Walking
		- Person Running
		- Person Squatting
		- Person Standing Up
		- Person Making/Assembling Stuffs with Hands (Hands Visible)
		- Person Batting Baseball
		(binary labels of 5 scene concepts)
		- Indoor Kitchen
		- Outdoor with Grass/Trees Visible
		- Baseball Field
		- Crowd (a group of 3+ people)
		- Cakes (close-up view)
		(binary labels of 10 audio concepts)
		- Outdoor Rural
		- Outdoor Urban
		- Indoor Quiet
		- Indoor Noisy
		- Original Audio
		- Dubbed Audio
		- Speech Comprehensible
		- Music
		- Cheering
		- Clapping


