Radical set determination for HMM based east asian character recognition
View Patent ↗Exemplary techniques are described for selecting radical sets for use in probabilistic East Asian character recognition algorithms. An exemplary technique includes applying a decomposition rule to each East Asian character of the set to generate a progressive splitting graph where the progressive splitting graph comprises radicals as nodes, formulating an optimization problem to find an optimal set of radicals to represent the set of East Asian characters using maximum likelihood and minimum description length and solving the optimization problem for the optimal set of radicals. Another exemplary technique includes selecting an optimal set of radicals by using a general function that characterizes a radical with respect to other East Asian characters and a complex function that characterizes complexity of a radical.
1. A method comprising:
executing processor-executable instructions on a processing unit, the processor-executable instructions comprising:
providing a set of East Asian characters;
applying a decomposition rule to each East Asian character of the set to generate a progressive splitting graph where the progressive splitting graph comprises radicals as nodes;
formulating an optimization problem to find an optimal set of radicals to represent the set of East Asian characters using a maximum likelihood and a minimum description length, the minimum description length comprising a penalty that depends on a number of parameters in a Hidden Markov Model based character recognition algorithm;
solving the optimization problem for the optimal set of radicals; and
using the optimal set of radicals in the Hidden Markov Model based character recognition algorithm.
2. The method of claim 1 wherein the formulating comprises defining layers of radicals and wherein the solving solves the optimization problem for an optimal layer.
3. The method of claim 1 wherein the formulating comprises applying a complete assertion condition.
4. The method of claim 1 wherein the formulating comprises defining layers of radicals and wherein the set of East Asian characters form a base radical layer.
5. The method of claim 1 wherein the formulating comprises defining layers of radicals and wherein simple strokes form a terminal radical layer.
6. The method of claim 1 wherein the formulating comprises defining at least three layers of radicals.
7. The method of claim 1 wherein the minimum description length comprises an additional penalty that depends on a number of training samples for training a probabilistic model based character recognition algorithm.
8. The method of claim 1 wherein the applying applies a decomposition rule to an East Asian character of the set and applies a decomposition rule to a radical generated by decomposition of an East Asian character of the set.
9. A computer-readable storage medium comprising the processor-executable instructions to perform the steps of claim 1 .
10. A method comprising:
executing processor-executable instructions on a processing unit, the processor-executable instructions comprising:
providing a set of East Asian characters;
for each East Asian character of the set, forming one or more sets of radicals and determining a value for each set of radicals based at least in part on
a general function that characterizes a radical with respect to other East Asian characters of the set of East Asian characters and a complex function that characterizes a complexity of the radical;
for each East Asian character, selecting the set of radicals that comprises the largest value;
determining an optimal set of radicals based on the selected sets of radicals for the set of East Asian characters; and
using the optimal set of radicals in a Hidden Markov Model based character recognition algorithm.
11. The method of claim 10 wherein the general function characterizes a radical according to a number of East Asian characters that include the radical.
12. The method of claim 10 wherein the complex function characterizes a radical according to a number of strokes in the radical.
13. The method of claim 10 wherein the determining determines a value for a radical set according to the function F:
F
(
R
i
)
=
(
C
′
(
r
a
)
·
C
′
(
r
b
)
·
C
′
(
r
c
)
·
…
·
C
′
(
r
ϕ
)
)
α
(
1
G
(
r
a
)
+
1
G
(
r
b
)
+
1
G
(
r
c
)
+
…
+
1
G
(
r
ϕ
)
)
β
·
ϕ
γ
where R, is a set of one or more radicals (r a , r b , . . . r φ ,), C′ is the complex function, G is the general function and α, β, γ are parameters.
14. The method of claim 13 wherein the parameters α, β, γ comprise adjustable weights.
15. A computer-readable storage medium comprising the processor-executable instructions to perform the steps of claim 10 .
16. A computing device comprising:
a processor;
a user input mechanism;
a display; and
control logic implemented at least in part by the processor to recognize an East Asian character based on a character recognition algorithm that uses a radical-based Hidden Markov Model, wherein the radical set comprises a radical set selected using a technique that uses a progressive splitting graph or a technique that uses a general function and a complex function, wherein the general function characterizes a radical with respect to other East Asian characters of a set of East Asian characters and wherein the complex function characterizes a complexity of the radical.
17. The computing device of claim 16 wherein the computing device comprises a cellular telephone.
18. The computing device of claim 16 wherein the computing device comprises a handheld computer.