Title :
Multivariate structural Bernoulli mixtures for recognition of handwritten numerals
Author :
Grim, JiYí ; Pudil, Pavel ; Somol, Petr
Author_Institution :
Inst. of Inf. Theory & Autom., Czechoslovak Acad. of Sci., Prague, Czech Republic
Abstract :
The structural optimization of a probabilistic neural network can be included into an expectation maximisation (EM) algorithm by introducing a special type of mixtures. The method has been applied to recognize unconstrained handwritten numerals from the database of Concordia University in Montreal. We discuss the possibility of a proper initialization of the EM algorithm for estimating the class-conditional multivariate Bernoulli mixtures
Keywords :
character recognition; maximum likelihood estimation; neural nets; optimisation; probability; class-conditional multivariate Bernoulli mixtures; expectation maximisation algorithm; multivariate structural Bernoulli mixtures; probabilistic neural network; structural optimization; unconstrained handwritten numerals; Automation; Databases; Handwriting recognition; Information theory; Input variables; Iterative algorithms; Neural networks; Neurons; Probability distribution; Structural engineering;
Conference_Titel :
Pattern Recognition, 2000. Proceedings. 15th International Conference on
Conference_Location :
Barcelona
Print_ISBN :
0-7695-0750-6
DOI :
10.1109/ICPR.2000.906142
