|Some time ago one teacher in Porto Alegre wrote about analysis of sign writing specific writings to tag by machine when two writings were similar enough to be identical. This effort became part of the Informatica department at the Catholic University of Pelotas. |
Humans can look at a glance and see how two writings are similar, it takes a machine quite a bit more testing and comparison to see a "critical difference" between signs like "MY" without movement, and "white" with movement.
. These two signs are radically different, but what about these three variants of "to" which essentially show the same movement.
The end position is the same. How do we classify and search for these? By the mathematical coding, the hand configurations of 2 and 3
will come out as closely related, the configuration of 1 would be different but could be at least in the same "family" of index fingers.
It is a challenge in search engines to identify identical or related "sign transcriptions" for the same movements.
[log in to unmask]
Clear writing moves business forward.
--- On Wed, 2/6/13, AGALYAVANEE <[log in to unmask]> wrote:
From: AGALYAVANEE <[log in to unmask]>
Subject: sign writing analysis
To: [log in to unmask]
Date: Wednesday, February 6, 2013, 1:22 PM
I would like to analyse sign writing based on user input. But I
am confused on how to identify and analyse the notations that intersect or
overlap. If anyone could provide me a hint, it would be of great help.