Data-driven design of HMM topology for online handwriting recognition

International Journal of Pattern Recognition and Artificial Intelligence

Volumn 15, Issue 1, 2001, Pages 107-121

  Lee, J J ^a   Kim, J ^a   Kim, J H ^a  

^a Korea Advanced Institute of Science and Technology (KAIST)   (South Korea)

Author keywords

Data driven topology design; Hangul recognition; Hidden Markov model; Multiple parallel path HMM; Online handwriting recognition; State tying based on structural similarity

Indexed keywords

COMPUTER AIDED LOGIC DESIGN; MARKOV PROCESSES; ONLINE SYSTEMS; PARALLEL PROCESSING SYSTEMS;

HANDWRITING RECOGNITION; HIDDEN MARKOV MODELS;

OPTICAL CHARACTER RECOGNITION;

EID: 0035247060     PISSN: 02180014     EISSN: None     Source Type: Journal    
DOI: 10.1142/S0218001401000769     Document Type: Article

References (18)

1. Recognition of unconstrained handwritten English words with character and ligature modeling
2. HMM based online handwriting recognition
3. Recognition of offline handwritten Korean characters
4. Hierarchical random graph representation of handwritten characters and its application to Hangul recognition
5. Recognition of online cursive Korean characters combining statistical and structural methods
6. Real-Time online unconstrained handwriting recognition using statistical methods
7. Recognition of isolated digits using hidden Markov models with continuous mixture densities
8. A segmental k-Means training procedure for connected word recognition
9. A comparison between continuous and discrete density hidden Markov models for cursive handwriting recognition
10. Ligature modeling for online cursive script recognition
11. A successive state splitting algorithm for efficient allophone modeling
12. Dynamic programming inference of Markov networks from finite sets of sample strings
13. The use of state tying in continuous speech recognition