PERSON DETECTION DEVICE AND PERSON DETECTION METHOD

• US 20130251203A1
• Filed: 10/05/2011
• Published: 09/26/2013
• Est. Priority Date: 12/09/2010
• Status: Active Grant

• Current Assignee: Panasonic Corporation (Panasonic Holdings Corporation)
• Original Assignee: C/O Panasonic Corporation
• Inventors: Tanabiki, Masamoto, Maruya, Kensuke, Fujita, Mitsuko, Kawaguchi, Kyoko, Sato, Yuji
• Granted Patent: US 8,934,674 B2
• Time in Patent Office: Days
• US Class Current: 382/103
• CPC Class Codes: G06T 2207/20096 Interactive definition of curve of interest
• CPC Class Codes: G06T 2207/30196 Human being; Person
• CPC Class Codes: G06T 7/75 involving models
• CPC Class Codes: G06V 40/10 Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
• CPC Class Codes: G06V 40/103 Static body considered as a whole, e.g. static pedestrian or occupant recognition

Abstract

Provided is a person detection device with which it is possible to estimate a state of a part of a person from an image. A person detection device (100) comprises: an evaluation unit (430) which acquires a prescribed outline of a person from an evaluation image; and a shoulder position calculation unit (440) and an orientation estimation unit (500) which estimate a state of a prescribed part of a person which is included in the evaluation image from the prescribed outline of the person which is acquired from the evaluation image, on the basis of an estimation model which denotes a relation between the prescribed outline and the state of the prescribed part of the person.

Claims

1-10. -10. (canceled)

11. A person detection apparatus, comprising:

an evaluation section that acquires an omega shape as an entire outline of a head and shoulders of a person from an evaluation image; and

an estimation section that estimates, on a basis of an estimation model that shows a relationship between the predetermined outline of a person and a state of a predetermined part of the person, a state of the predetermined part of a person included in the evaluation image based on the predetermined outline acquired from the evaluation image,

wherein;

the omega shape is defined by a positional relationship between a plurality of points forming the omega shape; and

the plurality of points include at least a point at an end portion of each of a left shoulder and a right shoulder and a point on each of a left side and a right side of a neck.

12. The person detection apparatus according to claim 11, wherein the positional relationship between the plurality of points is represented based on a distance or an angle with respect to a reference point that is set by a user.

13. The person detection apparatus according to claim 12, further comprising:

a feature generation section that acquires the omega shape of a person included in a sample image and a position of the shoulders of the person; and

the estimation model generation section that generates the estimation model based on the omega shape and the position of the shoulders acquired by the feature generation section.

14. The person detection apparatus according to claim 13, wherein the feature generation section displays the sample image, and receives a setting for the omega shape and a setting for a position of the shoulders from a user for each of the sample images.

15. The person detection apparatus according to claim 14, wherein the estimation model generation section generates the estimation model by performing a multiple regression analysis using information showing the omega shape as an explanatory variable and information showing a position of the shoulders as an objective variable.

16. The person detection apparatus according to claim 15, further comprising:

an omega generation section that generates, when a plurality of the omega shapes are used to generate the estimation model, one or a plurality of sample omega shapes based on the plurality of omega shapes, wherein

an evaluation section acquires the omega shape by matching between the sample omega shape and an edge portion of the evaluation image.

17. The person detection apparatus according to claim 16, wherein the omega generation section generates the plurality of sample omega shapes based on a result of a principal component analysis with respect to information showing the plurality of omega shapes and information showing a position of the shoulders.

18. A person detection method, comprising:

acquiring an omega shape as an entire outline of a head and shoulders of a person from an evaluation image; and

estimating, on a basis of an estimation model that shows a relationship between the predetermined outline of a person and a state of a predetermined part of the person, a state of the predetermined part of a person included in the evaluation image based on the predetermined outline acquired from the evaluation image;

wherein;

the omega shape is defined by a positional relationship between a plurality of points forming the omega shape; and

the plurality of points include at least a point at an end portion of each of a left shoulder and a right shoulder and a point on each of a left side and a right side of a neck.