Summaries home | UP
| email ]
Gary McGraw & Daniel Drasin. (1993) Recognition of Gridletters: Probing
the Behavior of Three Competing Models. In Proceedings of the Fifth Midwest
AI and Cognitive Science Conference, pages 63-67, April 1993.
Author of the summary: Patrawadee Prasangsit, 1999, email@example.com
The actual paper is online.
Cite this paper for:
Letter Spirit: A project that models aspects of human high-level
perception and creativity on a computer, focusing on the creative act of
DumRec, NetRec, FnetRec - the three models being compared
in the paper.
This paper compares the performance of three different models of letter
recognition in the Letter Spirit domain.
Categorical sameness is property possessed by instances of a
single letter in various styles (e.g., the letter 'a' in Times, Courier,
Stylistic sameness is property possessed by instances of various
letters in a single style (e.g., the letters 'a', 'b', and 'c' in Times).
Each letter is formed by a set of short line segments, called quanta,
on a fixed grid of dimension 3x7. See figure 2.
The three models are:
Associated with each training letter is a property list.
Given a mystery letter, DumRec computes its property list and compares
it with that of each training letter. The score is weighted sum of
the match of the property lists.
The weights play a crucial role in DumRec's performance. To modify
them is to "tune" DumRec.
2- or 3-layer feedforward connectionist networks trained using backpropogation.
56 input units, each corresponds to a quanta. 26 output units, each
corresponds to a letter of alphabet. Hidden layer may have 0-120
Major open problems: a learning rate for backpropogation and the number
of hidden units.
A variation of NetRec. Forces the network to pay more attention to
certain features as determined by human.
Train a number of small "subnets" to detect certain features. Examples
of features are height, weights, descenders, ascenders, different numbers
of tips, etc.
Input to the letter-recognizer network is the existing 56 input units as
well as outputs from the subnets.
For all models, the performance is still unacceptable (too many mis-categorizations).
The reason could be
The percentage of successful recognition: DumRec 74.3%, FnetRec 72.84%,
DumRec performs best among all, however the differences are not large.
DumRec, most of the time, guess the correct letter or a "reasonable" wrong
NetRec and FnetRec are very similar when compare letter by letter, though
in general the latter slightly outperforms.
DumRec - Probably because the features considered are too low-level.
Better recognition requires the use of higher-level features (e.g., roles).
NetRec and FnetRec - Style may interfere with recognition.
Summary author's notes:
to the Cognitive Science Summaries homepage
Cognitive Science Summaries Webmaster:
JimDavies ( firstname.lastname@example.org
Last modified: Thu May 6 09:02:17 EDT 1999