JP2018045302A5

JP2018045302A5 -

Info

Publication number: JP2018045302A5
Application number: JP2016177789A
Authority: JP
Filing date: 2016-09-12
Publication date: 2019-10-24
Anticipated expiration: 2036-09-12

Description

本発明に係る情報処理装置は、異なる条件で取得した複数のデータ群を、前記複数のデータ群に係る個々のデータの特徴に基づいてそれぞれ複数のグループに分割する分割手段と、前記分割手段によって分割されたグループ間の統合適正度を算出する算出手段と、前記算出手段によって算出された統合適正度に基づいて、前記分割されたグループを統合データグループとして統合する統合手段と、前記統合手段によって統合された統合データグループからデータを抽出して、認識用の辞書を生成するための学習処理用のデータのセットを設定する設定手段と、を有することを特徴とする。 An information processing apparatus according to the present invention includes a dividing unit configured to divide a plurality of data groups acquired under different conditions into a plurality of groups based on characteristics of individual data related to the plurality of data groups, and the dividing unit. A calculation unit that calculates an appropriateness of integration between the divided groups; an integration unit that integrates the divided groups as an integrated data group based on the integrated appropriateness calculated by the calculation unit; and the integration unit And setting means for setting a set of data for learning processing for extracting data from the integrated integrated data group and generating a dictionary for recognition.

Claims

Dividing means for dividing a plurality of data groups acquired under different conditions into a plurality of groups based on the characteristics of the individual data related to the plurality of data groups,
Calculating means for calculating the degree of integration appropriate between the groups divided by the dividing means;
An integration unit that integrates the divided groups as an integrated data group based on the integration suitability calculated by the calculation unit;
Setting means for extracting data from the integrated data group integrated by the integrating means and setting a set of data for learning processing for generating a dictionary for recognition;
An information processing apparatus comprising:

The information processing apparatus according to claim 1, wherein the setting unit extracts some data from the integrated data group when the number of data in the integrated data group exceeds a predetermined value.

The information processing apparatus according to claim 2, wherein the setting unit determines data to be extracted from the integrated data group based on a transfer cost for transferring the learning processing data.

The information processing apparatus according to claim 1, wherein the calculation unit calculates a similarity between the divided groups as the integration appropriateness.

The information processing apparatus according to claim 4, wherein the calculation unit further calculates the integration appropriateness based on attribute information relating to the plurality of data groups.

The information processing apparatus according to claim 1, wherein the plurality of data groups are feature amounts of images acquired from a plurality of cameras, respectively.

The dividing unit divides the plurality of groups by generating a dictionary by performing a learning process using the data group,
The information processing apparatus according to claim 6, wherein the calculation unit calculates the integration appropriateness by performing recognition processing based on the generated dictionary.

Generating means for performing learning processing using a set of data for learning processing set by the setting means, and generating a dictionary related to the set;
Verification means for verifying the dictionary generated by the generation means;
The information processing apparatus according to claim 1, further comprising:

A dividing step of dividing a plurality of data groups acquired under different conditions into a plurality of groups based on the characteristics of individual data related to the plurality of data groups,
A calculation step of calculating an integration suitability between the groups divided in the division step;
An integration step of integrating the divided groups as an integrated data group based on the integration suitability calculated in the calculation step;
A setting step of extracting data from the integrated data group integrated in the integration step and setting a set of data for learning processing for generating a dictionary for recognition;
An information processing method characterized by comprising:

The program for functioning a computer as each means of the information processing apparatus of any one of Claims 1-8.