HEAD DETECTING APPARATUS, HEAD DETECTING METHOD, AND HEAD DETECTING PROGRAM

• US 20090022398A1
• Filed: 09/18/2008
• Published: 01/22/2009
• Est. Priority Date: 12/16/2005
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Maeda, Junji, Ishikawa, Hiroshi, Nakano, Hiroki
• Granted Patent: US 7,720,285 B2
• Time in Patent Office: Days
• US Class Current: 382/181
• CPC Class Codes: G06V 40/164 using holistic features

Abstract

A head detecting apparatus, including a foreground extraction section for extracting a foreground region in which a person is captured from an input image; a first main axis computing section which includes a first moment computing section for computing a moment around a center of gravity of the foreground region and calculating a main axis of the foreground region based on the moment around the center of gravity of the foreground region; a head computing section for computing a head region included in the foreground region as a part thereof based on the main axis of the foreground region and a shape of the foreground region; and an ellipse determining section for determining an ellipse to be applied to a person'"'"'s head based on a shape of the head region.

Claims

1. A head detecting method comprising the steps of:

a foreground extracting step for extracting a foreground region, in which a person has been taken, from an input image;

a first main axis computing step for computing a main axis of the foreground region, said first main axis computing step including a first moment computing step for computing a moment around a center of gravity of the foreground region, and calculating the main axis of the foreground region based on the moment around the center of gravity of the foreground region;

a head region computing step for computing a head region included in the foreground region as a part thereof, based on the main axis of the foreground region and a shape of the foreground region;

an ellipse determining step for determining an ellipse to be applied to the head of the person, based on a shape of the head region; and

an input step for inputting a plurality of frame images constituting a moving image as a plurality of the input images;

and wherein

the first main axis computing step includes;

a first orthogonal axes computing step for computing two orthogonal axes in the foreground region, based on the moment around the center of gravity of the foreground region; and

a first main axis selection step for selecting one axis in the two orthogonal axes calculated by the first orthogonal axes computing step, which forms smaller angle with the main axis of the foreground region calculated in at least one of a previous frame and a next frame, as the main axis of the foreground region.

2. A head detecting method comprising:

a foreground extracting step for extracting a foreground region, in which a person has been taken, from an input image;

a first main axis computing step for computing a main axis of the foreground region;

a head region computing step for computing a head region included in the foreground region as a part thereof, based on the main axis of the foreground region and a shape of the foreground region;

an ellipse determining step for determining an ellipse to be applied to the head of the person, based on a shape of the head region;

a search step for searching a top point of a head of a person, by setting a starting point at one of intersecting points of the outline of the foreground region and the main axis, which is deemed to be at a higher position in the image pick-up target space, and following pixels of the outline in the foreground region; and

a circle computing step for computing a circle, in which the proportion of pixels of the foreground region included in the circle is equal to or greater than a predetermined reference proportion, as a circle surrounding the head, among circles which have centers on a line connecting the top point and the center of gravity of the foreground region and pass through the top point of the head;

wherein the foreground region included in the circle surrounding the head is calculated as the head region.

3. The head detecting method, according to claim 2, wherein

the search step follows an outline of the foreground region from the starting point in a direction away from the center of gravity of the foreground region, and sets a point where the distance from the center of gravity is maximum as the top point of the head.

4. The head detecting method, according to claim 2, wherein the search step determines the top point of the head under a condition that an angle formed by a central line which connects the top point of the head and the center of gravity of the foreground region, and the main axis of the foreground region is less than a predetermined reference amount of degree.

5. The head detecting method, according to claim 2, wherein:

the search step determines the top point of the head under a condition that a length of the outline between the starting point and the top point of the head is equal to or less than a predetermined reference value.

6. The head detecting method, according to claim 2, wherein:

the circle computing step corrects the circle surrounding the head to enlarge it so that the circle includes a whole flesh color region under a condition that the circle, surrounding the head calculated so that the proportion of pixels in the foreground region included in the circle is equal to or greater than a predetermined reference proportion, crosses the flesh color region of the foreground region.

7. A head detecting method comprising the steps of:

a foreground extracting step for extracting a foreground region, in which a person has been taken, from an input image;

a first main axis computing step for computing a main axis of the foreground region;

a head region computing step for computing a head region included in the foreground region as a part thereof, based on the main axis of the foreground region and a shape of the foreground region; and

an ellipse determining step for determining an ellipse to be applied to the head of the person, based on a shape of the head region, wherein

the ellipse detection step includes a second principal computing step for computing a main axis of the head region, and determining the ellipse having the main axis as a long axis and wherein

the second main axis computing step includes;

a second moment computing step for computing a moment around a center of gravity of the head region;

a second orthogonal axes computing step for computing two orthogonal axes in the head region based on the moment around the center of gravity of the head region; and

a second main axis selection step for selecting an axis which forms smaller angle with the main axis of the foreground region, in the two orthogonal axes.

8. A head detecting method comprising the steps of:

a foreground extracting step for extracting a foreground region, in which a person has been taken, from an input image;

a first main axis computing step for computing a main axis of the foreground region;

a head region computing step for computing a head region included in the foreground region as a part thereof, based on the main axis of the foreground region and a shape of the foreground region; and

an ellipse determining step for determining an ellipse to be applied to the head of the person, based on a shape of the head region, wherein;

the ellipse detection step includes a second principal computing step for computing a main axis of the head region, and determines the ellipse having the main axis as a long axis, a flesh color extraction step for extracting a biggest flesh color area in the head region;

a short axis determination step for determining a length of the short axis according to a largest distance between the main axis of the head region and pixels in the flesh color area and a long axis determination step for setting a value obtained by multiplying the length of the short axis of the ellipse determined by the short axis determination step by a predetermined value which is larger than 1, as a length of a long axis of the ellipse.

9. A head detecting method comprising:

a foreground extracting step for extracting a foreground region, in which a person has been taken, from an input image;

a first main axis computing step for computing a main axis of the foreground region;

a head region computing step for computing a head region included in the foreground region as a part thereof, based on the main axis of the foreground region and a shape of the foreground region;

an ellipse determining step for determining an ellipse to be applied to the head of the person, based on a shape of the head region;

an input step for inputting a plurality of frame images constituting a moving image as a plurality of the input image; and

a detection step for detecting a change of a position and an orientation of the person'"'"'s head based on a change of the ellipse determined according to the plurality of frame images, wherein;

the input step inputs the plurality of the frame images constituting the moving image captured by the image pick-up device for a monitoring purpose;

and further including;

a storage step for storing history of the change of the position and the orientation of the person'"'"'s head for each of a plurality of persons in a storage section; and

an abnormality detection step for detecting an abnormal action of the person under a condition that a difference between an amount of the change of the position and the orientation of the person'"'"'s head and a reference amount of change based on a position and a change of other person'"'"'s head stored in the storage section is equal to or more than a predetermined amount.