NON-TRANSITORY COMPUTER READABLE MEDIUM, ACTION DETECTION APPARATUS, AND ACTION DETECTION METHOD

• US 20140039827A1
• Filed: 02/15/2013
• Published: 02/06/2014
• Est. Priority Date: 08/06/2012
• Status: Active Grant

• Current Assignee: Fujifilm Business Innovation Corp. (Fujifilm Holdings Corporation)
• Original Assignee: Fuji Xerox Company Limited (Xerox Holdings Corp.)
• Inventors: YUZAWA, Hideto
• Granted Patent: US 9,638,713 B2
• Time in Patent Office: Days
• US Class Current: 702/141
• CPC Class Codes: A61B 2562/0219 Inertial sensors, e.g. accelerometers, gyroscopes, tilt switches
• CPC Class Codes: G01P 13/00 Indicating or recording presence, absence, or direction, of movement electric switches H01H; counting moving objects G06M7/00
• CPC Class Codes: G01P 15/00 Measuring acceleration; Measuring deceleration; Measuring shock, i.e. sudden change of acceleration
• CPC Class Codes: G01P 15/0891 with indication of predetermined acceleration values G01P15/135 takes precedence
• CPC Class Codes: G01P 15/18 in two or more dimensions

Abstract

A non-transitory computer readable medium stores a program causing a computer to execute a process. The process includes calculating a variance of acceleration in individual sections obtained by dividing, by a first predetermined time interval, acceleration information for multiple axis directions detected from a subject whose action is to be determined; extracting, as a first candidate region, a section with the largest variance of acceleration from among the sections; calculating a variance of acceleration in individual sections obtained by shifting the first candidate region forward and backward by a second predetermined time interval; selecting, as a second candidate region, a section with the largest variance of acceleration from among the first candidate region and the sections obtained by shifting the first candidate region forward and backward; calculating a feature value for the second candidate region; and determining the action of the subject based on the feature value.

Claims

1. A non-transitory computer readable medium storing a program causing a computer to execute a process for detecting an action, the process comprising:

calculating a variance of acceleration in individual sections obtained by dividing pieces of acceleration information for a plurality of axis directions by a first predetermined time interval, the pieces of acceleration information being detected from a subject whose action involving a movement is to be determined;

extracting, as a first candidate region, a section with the largest variance of acceleration from among the sections;

calculating a variance of acceleration in individual sections obtained by shifting the first candidate region forward and backward by a second predetermined time interval;

selecting, as a second candidate region, a section with the largest variance of acceleration from among the first candidate region, the section obtained by shifting the first candidate region forward, and the section obtained by shifting the first candidate region backward;

calculating a feature value for the selected second candidate region; and

determining the action of the subject on the basis of the calculated feature value.

2. The non-transitory computer readable medium according to claim 1, wherein in the determining, the action of the subject is determined using a decision tree regarding the calculated feature value.

3. The non-transitory computer readable medium according to claim 2, wherein in the determining, noise is determined using a component of the feature value corresponding to a lateral direction of the subject, and a nodding action is the action of the subject to be determined.

4. The non-transitory computer readable medium according to claim 1, wherein the pieces of acceleration information are detected by an acceleration sensor that is attached in the vicinity of the neck of a person serving as the subject.

5. The non-transitory computer readable medium according to claim 2, wherein the pieces of acceleration information are detected by an acceleration sensor that is attached in the vicinity of the neck of a person serving as the subject.

6. The non-transitory computer readable medium according to claim 3, wherein the pieces of acceleration information are detected by an acceleration sensor that is attached in the vicinity of the neck of a person serving as the subject.

7. An action detection apparatus comprising:

a candidate region extracting unit that calculates a variance of acceleration in individual sections obtained by dividing pieces of acceleration information for a plurality of axis directions by a first predetermined time interval, the pieces of acceleration information being detected from a subject whose action involving a movement is to be determined, and that extracts, as a first candidate region, a section with the largest variance of acceleration from among the sections;

a candidate region selecting unit that calculates a variance of acceleration in individual sections obtained by shifting the first candidate region forward and backward by a second predetermined time interval, and that selects, as a second candidate region, a section with the largest variance of acceleration from among the first candidate region, the section obtained by shifting the first candidate region forward, and the section obtained by shifting the first candidate region backward;

a feature value calculating unit that calculates a feature value for the second candidate region selected by the candidate region selecting unit; and

a determining unit that determines the action of the subject on the basis of the feature value calculated by the feature value calculating unit.

8. An action detection method comprising:

calculating a variance of acceleration in individual sections obtained by dividing pieces of acceleration information for a plurality of axis directions by a first predetermined time interval, the pieces of acceleration information being detected from a subject whose action involving a movement is to be determined;

extracting, as a first candidate region, a section with the largest variance of acceleration from among the sections;

calculating a variance of acceleration in individual sections obtained by shifting the first candidate region forward and backward by a second predetermined time interval;

selecting, as a second candidate region, a section with the largest variance of acceleration from among the first candidate region, the section obtained by shifting the first candidate region forward, and the section obtained by shifting the first candidate region backward;

calculating a feature value for the selected second candidate region; and

determining the action of the subject on the basis of the calculated feature value.