Pattern recognition system

Claim 1

A pattern recognition system characterized by comprising: a feature extraction section (1) for extracting feature data of an input pattern; a memory (3) storing a reference pattern of each category; a multiple similarity calculation section (2) for calculating multiple similarities between the feature data extracted by said feature extraction section and the reference patterns stored in said memory; and posterior probability conversion means for converting the multiple similarities obtained by said similarity calculation section into posterior probabilities, wherein said posterior probability conversion means (4) includes: a parameter table (5) storing parameter sets each composed of a plurality of paramenter pairs of α and β, said parameter α indicating a number of com­ponents which are not involved in a subspace spanned by a reference pattern, the parameter β indicating a variance in a direction of non-reference subspace, and both parameters α and β being estimated from multiple similarity values between a reference pattern and training data, for each category of learning patterns, and for every category of reference patterns; and means (4) for calculating the posterior probabilities based on the multiple similarities calculated by said multiple similarity calculation section and the parameter sets stored in said parameter table.