Feature transformation apparatus and feature transformation method

A feature transformation apparatus comprising:

• a prescribed activity data acquisition section configured to acquire respective prescribed activity data of one or more reference individuals and a subject when the one or more reference individuals and the subject perform a prescribed activity;
• a first feature extraction section configured to extract a reference individual feature of the prescribed activity data of the one or more reference individuals and a subject feature of the prescribed activity data of the subject from the respective prescribed activity data;
• a standard feature calculation section configured to calculate at least one standard feature based on the one or more reference individual features;
• a transform matrix calculation section configured to calculate a transform matrix between the calculated standard feature and the subject feature;
• a recognition object activity data acquisition section configured to acquire recognition object activity data when the subject performs a recognition object activity;
• a second feature extraction section configured to extract a recognition object feature of the recognition object activity data of the subject from the recognition object activity data of the subject;
• and
• a transformation section configured to transform the recognition object feature of the recognition object activity data using the transform matrix.

A feature transformation method to be executed by an information processing apparatus including a control unit and a storage unit, the feature transformation method comprising:

• the control unit acquiring respective prescribed activity data of one or more reference individuals and a subject when the one or more reference individuals and the subject perform a prescribed activity and storing the respective prescribed activity data in the storage unit;
• the control unit reading out respective prescribed activity data stored in the storage unit and extracting a reference individual feature of the prescribed activity data of the one or more reference individuals and a subject feature of the prescribed activity data of the subject from the respective read-out prescribed activity data;
• the control unit calculating at least one standard feature based on the one or more reference individual features;
• the control unit calculating a transform matrix between the calculated standard feature and the subject feature and storing the transform matrix in the storage unit;
• the control unit acquiring recognition object activity data when the subject performs a recognition object activity and storing the recognition object activity data in the storage unit;
• the control unit reading out the recognition object activity data of the subject stored in the storage unit and extracting a recognition object feature of the recognition object activity data of the subject from the read-out recognition object activity data;
• and
• the control unit reading out the transform matrix stored in the storage unit and transforming the recognition object feature of the recognition object activity data using the read-out transform matrix.