File(type)s:

- "collection.xml":  video id  is linked to the  video file name and its
	   	     use for development or test.

   2. Jul 2003 	8 files were withdrawn from the test collection due to
		significant encoding problems over a large portion of
		each video. Such videos are marked "withdrawn.from.test"
  

- "xxx.mp7.xml"  : Mpeg-7 output, where "xxx" is the video id instead of the
  video file name.



A few small points that are related to the Mpeg-7 output:

- The media time format is based on the Gregorian day time (ISO 8601)
norm.  Fractions are defined by counting pre-sepcified fractions of a
second.  In our case, the frame rate is 29.97.  One fraction of a
second is thus specified as "PT1001N30000F".

- Trec video id has the format of "XXX" and shot id "shotXXX_YYY".
The "XXX" is the sequence number of video onto
which the video file name is mapped, this is based on the
"collection.xml" file. The "YYY" is the sequence number of the
shot. 

- There are two types of keyframes : one "representative keyframe" (with
"RKF" in the subshot name) corresponding to the main (or the single)
subshot within the shot, zero, one or more additional "non
representative keyframe" (with "NRKF" in the shot name) corresponding
to other subshots (if any) within the shot. Subshots within a shots
correspond to extracted shots merged together in order to enforce the
2s minimum duration of a shot.


Information about the segmentation:

- The segmentation was performed using an improved version of the the
system used by CLIPS-IMAG for the TREC 2002 shot boundary
determination task.

- The segmentation has been done so that no shot has a duration of less
than 2 second (or 60 frames) so that assessors can confidently
evaluate them. In order to satisfy this constraint, short shots have
been merged with their neighbors. Therefore, many shots (roughly 20% I
would say) actually contain several subshots. When a shot contains
several subshots (and we know that because we know the segmentation
before and after the merging od short shots), the corresponding
keyframe is always chosen within the longest subshot.

- The segmentation is such that there are no gaps between shots and no
shots overlap. When a gradual transition is detected, the end and the
begining of the previous and next shots are set to the middle point of
the transition (i.e. shots begins and ends in the middle of the
transitions). Again, although the transition boundary information is
lost in the final segmentation, it it used for keyframe extraction and
the keyframe is never within a portion of a (detected) gradual
transition.