RE: trecvid - additional known items
- Subject: RE: trecvid - additional known items
- From: Alexander Hauptmann <email@example.com>
- Date: Mon, 20 Aug 2001 17:16:07 -0400
- Content-Transfer-Encoding: 7bit
- Content-Type: text/plain; charset="us-ascii"
- Importance: Normal
- In-Reply-To: <3B811398.9CE155C4@nist.gov>
- MMDF-Warning: Parse error in original version of preceding line at speech1.cs.cmu.edu
I am okay with keeping the known items as they are.
I believe there may be other known items for topic 70 that we didn't get
using our interactive system,
as I would expect every source mpg file to list a producer or production
Looking at the proposed known items more closely, there were some instances
we found several contiguous individual "shots" that were grouped together in
original known items, which would result in very different scoring.
I guess next year the shot as a unit of retrieval should be clarified to
stricly one shot per result or to combine adjacent shots automatically.
Also, it was not clear if e.g. a topic looking for water skiers should find
frames inside a shot where the skier appeared until the skier disappears, or
if the whole shot which includes the skier somewhere in the middle would be
acceptable as an answer.
If we had been on the ball, we would have asked these questions much
From: firstname.lastname@example.org [mailto:email@example.com]On Behalf Of Paul Over
Sent: Monday, August 20, 2001 9:49 AM
To: Multiple recipients of list
Subject: Re: trecvid - additional known items
Thanks for the additional known-item candidates. I take it
you are proposing all the search result items be known-items?
I agree, this will take some time to work through.
Paul Over - Retrieval Group
Information Access Division
Information Technology Laboratory
National Institute of Standards and Technology
Bldg. 225 Rm. A211 (Mailstop 8940)
Gaithersburg, MD 20899-8940 USA
Voice: 301 975-6784 Fax: 301 975-5287
Date Index |
Thread Index |
Problems or questions? Contact firstname.lastname@example.org