Standard model creating device and standard model creating method

Abstract

The standard model creating apparatus which provides a high-precision standard model used for: pattern recognition such as speech recognition, character recognition, or image recognition using a probability model based on a hidden Markov model, Bayesian theory, or linear discrimination analysis; intention interpretation using a probability model such as a Bayesian net; data-mining performed using a probability model; and so forth, the apparatus comprising: a reference model preparing unit (102) operable to prepare at least one reference model; a reference model storing unit (103) operable to store the reference model (121) prepared by the reference model preparing unit (102); and a standard model creating unit (104) operable to create a standard model (122) by calculating statistics of the standard model so as to maximize or locally maximize the probability or likelihood with respect to the at least one reference model stored in the reference model storing unit (103).

Description

technical field [0001] The invention relates to a standard model production device and method thereof, which are used for pattern recognition such as sound recognition, character recognition, and image recognition based on probability models such as hidden Markov model, Bayesian logic, and linear judgment. Intention understanding (recognition of intent) of network equiprobability model, data acquisition based on probability model (recognition of data characteristics), person detection based on probability model, fingerprint authentication, face authentication, iridescent authentication (identify object, judge whether it is a specific object ), stock forecasts, weather forecasts, etc. (judgment after recognizing the situation), synthesis of voices of multiple speakers, synthesis of multiple facial images, etc. (people recognize the synthesized model interestingly), etc. Background technique [0002] In recent years, with the popularization of the Internet and the like, the in...

AI Technical Summary

Problems solved by technology

[0054] In the fifth conventional method, although the standard model is produced by synthesizing Gaussian distributions having the same serial number, in order to produce the best standard model, the generally synthesized Gaussian distributions are not limited to one-to-one correspondence, so the recognition accuracy is low

In addition, standard models cannot be produced when multiple reference models have different mixture numbers

In addition, the Gaussian distribution in the reference model is generally not assigned consecutive numbers, and the standard model cannot be created in this case

In addition, it is not possible to control the mixture number of the standard model according to the specification

Images