[evaal contest] [localisation] regarding phase3: the disturber

[Francesco Potortì]

Dear all,

this is in response to competitors who asked about the possibility of
sending more than one position to the system (one about the actor, one
more about the disturber).  This would not be acceptable: the rationale
of the third phase of the competition is in fact to check the system's
ability to identify the actor without being fooled by the presence of
another person.

This should be much easier for systems that physically tag the actor.
Those systems, however, will likely have a lower "acceptance" score,
because of the need of phisically wearing something.  Moreover, the
advantage is on one out of the three phases, and only as far as the
accuracy score is concerned.  If you look at the way the scores are
computed, you see that the advantage is roughly on 33% of 25% of the
total score.

Systems that do not tag the actor may benefit from tracking.  Tracking
can be helped by the fact that the actor starts moving 5 seconds before
the disturber, and by the use of contextual events, knowing that the
actor will be the first to generate a contextual event.

Systems that do not physically tag the actor may be fooled if the actor
and the disturber come nearby.  That is their main disadvantage with
respect to systems that tag the actor.  Systems that are smart enough
can use a (maybe risky) trick in case they are undecided among two
paths: track the first such path on the first test and the second such
path on the second test.  Since the best test only is considered, this
may bring an advantage.


-- 
Francesco Potortì (ricercatore)        Voice:  +39.050.315.3058 (op.2111)
ISTI - Area della ricerca CNR          Mobile: +39.348.8283.107
via G. Moruzzi 1, I-56124 Pisa         Fax:    +39.050.315.2040	 
(entrance 20, 1st floor, room C71)     Web:    http://fly.isti.cnr.it