Image processing device, method, and program

• US 7,796,786 B2
• Filed: 11/15/2006
• Issued: 09/14/2010
• Est. Priority Date: 12/02/2005
• Status: Expired due to Fees

• Current Assignee: Omron Corporation
• Original Assignee: Omron Corporation
• Inventors: Yamaguchi, Yoshinori, Kimura, Masataka, Ishikawa, Takamitsu, Sugiura, Mitsunori, Kakinuma, Takeshi
• Primary Examiner(s): Bella; Matthew C
• Assistant Examiner(s): RAHMJOO, MANUCHEHR
• Application Number:
  US11/599,827
• Publication Number:
  US 20070127788A1
• Time in Patent Office: 1,399 Days
• Field of Search: 382/118, 382/123, 382/151, 382/159, 382/190
• US Class Current: 382/118
• CPC Class Codes: G06V 40/161 Detection; Localisation; Normalisation

Abstract

To prevent detection of facial images for whom detection is unnecessary, without reducing the detection ratio of registered players from facial images.

An image acquisition unit acquires an image taken by a camera, and provides the image to a masking process unit. The masking process unit reads a masking pattern that has already been adjusted by a masking pattern adjustment unit and a movable masking area setting unit so that characteristic quantities of a facial image that does not need to be recognized as a facial image cannot be detected, from a masking pattern recording unit. A masking process is applied to the image provided, and the masked image is provided to a facial image extraction unit. The facial image extraction unit extracts a rectangular shaped facial image, and provides the facial image to a transmission unit. The transmission unit transmits the rectangular facial image provided to a face recognition server.

Claims

1. An image processing device, comprising:

imaging means that takes an image;

facial image extraction means that extracts a facial image from the image;

masking pattern generation means that generates a masking pattern to mask a part of an image, and repeatedly changes the masking pattern until a state where a facial image other than the facial image of a visitor cannot be extracted from the image by the facial image extraction means; and

recording means that records the masking pattern when the facial image extraction means cannot extract the facial image other than the facial image of a visitor from the image that has been masked using the masking pattern generated by the masking pattern generation means.

2. The image processing device according to claim 1, wherein

the masking pattern generation means further comprises sub-division means that sub-divides the facial image into predetermined sub-division areas, and

the masking pattern generation means generates masking patterns that mask a part of the facial image using units of the sub-divided areas sub-divided by the sub-division means, and the position and size of the masking pattern is repeatedly changed in units of the sub-divided areas sub-divided by the sub-division means until a state where a facial image other than the facial image of a visitor cannot be extracted from the image by the facial image extraction means.

3. The image processing device according to claim 2, wherein the predetermined sub-division areas include areas of the sub-divided image in a scan line form, or areas of the sub-divided image in a matrix faun.

4. The image processing device according to claim 3, wherein if the predetermined sub-division areas are areas sub-divided in a matrix form, the masking pattern generation means generates the masking pattern for masking by changing the size of part of the sub-divided area while successively moving the facial image in a whirlpool shape from an area in approximately the center of a facial image other than the facial image of a visitor.

5. The image processing device according to claim 1, wherein the facial image other than the facial image of a visitor is a facial image extracted by the facial image extraction means from an image taken by the imaging means under conditions in which no visitors are present.

6. The image processing device according to claim 1, further comprising:

detection means that detects positions of movement of movable equipment; and

tracing means that traces the masking pattern generated by the masking pattern generation means in a state where the facial image generation means cannot extract a facial image from the image taken by the imaging means, based on the positions of movement detected by the detection means,

wherein the masking pattern recording means records the masking pattern including a range that the masking pattern has been traced by the tracing means.

7. A surveillance device, comprising:

accumulation means that accumulates facial images of registered persons;

masking means that masks the images using masking patterns recorded in the recording means according to claim 1;

degree of similarity calculation means that calculates the degree of similarity between the facial images extracted by the facial image extraction means and the facial images of registered persons accumulated in the accumulation means, based on characteristic quantities for recognizing the facial images from the facial images extracted by the facial image extraction means from the images masked by the masking means; and

determination means that determines whether a registered person is present or not based on the degree of similarity calculated by the degree of similarity calculation means.

8. An image processing method, comprising:

an imaging step of taking an image;

a facial image extraction step of extracting a facial image from the image;

a masking pattern generation step of generating a masking pattern for masking a part of the image, and repeatedly changing the masking pattern until a state where a facial image other than the facial image of a visitor cannot be extracted from the image, by the process in the facial image extraction step; and

a recording step of recording the masking pattern when the facial image extraction means cannot extract a facial image other than the facial image of a visitor from the image that has been masked using the masking pattern generated in the process of the masking pattern generation step.

9. A program stored on non-transitory computer readable medium, comprising:

an imaging step of taking an image;

a facial image extraction step of extracting a facial image from the image;

a masking pattern generation step of generating a masking pattern for masking apart of the image, and repeatedly changing the masking pattern until a state where a facial image other than the facial image of a visitor cannot be extracted from the image by the process in the facial image extraction step; and

a recording step of recording the masking pattern when the facial image extraction means cannot extract a facial image other than the facial image of a visitor from the image that has been masked using the masking pattern generated in the process of the masking pattern generation step.