===========================================================
Hopkins 155 Additional Sequences: Outliers and Missing Data
===========================================================
Version: April 19th, 2010

Additional sequences description
--------------------------------
These additional sequences contain various kinds of moving objects. All sequences contain from three to five objects undergoing a variety of motions (planar, rigid, non-rigid, etc.). As opposed to the ones contained in the Hopkins 155 ([1], [2]), these sequences contain both missing points (frames for which the tracked point becomes occluded and cannot be followed by the tracker) and outliers (tracks that do not correspond to any one of the groups). These sequences have been used in [3] and [4].

Database content
----------------

# of sequences 		4
Avg. # of groups 	4.0
Avg. # of points 	975
Avg. # of frames 	54
Avg. % of outliers 	13.92%
Avg. % of missing data 	29.48%

File format
-----------
The file format is analogous to the one used in the Hopkins 155 ([1]). The file *_truth.mat, however, contains a the new variable m, with the information on the frames for which the points are missing:

. m: a matrix PxF (P = number of points, F = number of frames) where the p,f-th entry is 1 if the point p is visible in the frame f and zero otherwise.

In addition, the vector s (which in the main Hopkins 155 contains the ground-truth segmentation) as a zero for the p-th entry if the p-th track is an outlier.

References
----------

[1] http://www.vision.jhu.edu/data/hopkins155/

[2] R. Tron and R. Vidal.
A Benchmark for the Comparison of 3-D Motion Segmentation Algorithms.
IEEE International Conference on Computer Vision and Pattern Recognition, 2007. 

[3] S. Rao, R. Tron, Y. Ma, and R. Vidal.
Motion Segmentation via Robust Subspace Separation in the Presence of Outlying, Incomplete, or Corrupted Trajectories.
IEEE International Conference on Computer Vision and Pattern Recognition, 2008. 

[4] S. Rao, R. Tron, R. Vidal, and Y. Ma.
Motion Segmentation in the Presence of Outlying, Incomplete, or Corrupted Trajectories.
IEEE Transactions on Pattern Analysis and Machine Intelligence, 2009. 

Copyright 2010 VisionLab
