Face identification device and face identification method

• US 20040081338A1
• Filed: 07/29/2003
• Published: 04/29/2004
• Est. Priority Date: 07/30/2002
• Status: Active Grant

• Current Assignee: Omron Corporation
• Original Assignee: Omron Corporation
• Inventors: Takenaka, Hideki
• Granted Patent: US 7,440,594 B2
• Time in Patent Office: Days
• US Class Current: 382/118
• CPC Class Codes: G06T 2207/30201 Face
• CPC Class Codes: G06T 5/70 Denoising; Smoothing
• CPC Class Codes: G06V 40/16 Human faces, e.g. facial parts, sketches or expressions
• CPC Class Codes: G08B 13/1961 Movement detection not involving frame subtraction, e.g. motion detection on the basis of luminance changes in the image
• CPC Class Codes: G08B 13/19686 Interfaces masking personal details for privacy, e.g. blurring faces, vehicle license plates

Abstract

The present invention is to protect the right of portrait or privacy of people other than a specific person, and to enable easy recognition of information other than the face of the specific person and the situation around the specific person. Images videotaped by the surveillance cameras are captured in the computer, face images are detected therefrom, and the detected face images are compared with the face image of a specific person registered in the storage device. As the results of the comparison, when these face images do not match with each other, a mosaic process is applied exclusively to the detected face images to protect the privacy of the people other than the specific person, whereas the mosaic process is not applied to the parts other than the faces, such as dress and baggage, thereby keeping them as they are. When these face images match with each other, the face image of the specific person is not applied with the mosaic process, thereby keeping his or her face and remaining parts as they are.

Claims

1. A face identification device comprising:

detection means for detecting face images from human body images taken by a camera;

storage means in which a face image of a specific person is previously stored;

determination means for determining whether a face image detected by said detection means matches with the face image stored in said storage means by comparing both face images; and

abstraction means for applying an abstraction process to a predetermined face image out of the face images detected by said detection means in order to make the predetermined face image unrecognizable,

said abstraction means applying the abstraction process exclusively to a detected face image when said determination means determines that both face images do not match with each other, and not applying the abstraction process to a detected face image when said determination means determines that both face images match with each other.

2. The face identification device according to claim 1, wherein said abstraction process is a mosaic process for making a face image portion mosaic.

3. The face identification device according to claim 1 or 2, wherein when said determination means determines that both face images match with each other, a detected face image is not applied with the abstraction process and is applied with a marker.

4. The face identification device according to any one of claims 1 to 3, wherein when a face image detected by said detection means is determined to match with the face image stored in said storage means, the image of the specific person which is not applied with the abstraction process on the face and remaining parts thereof and the images of people other than the specific person which are applied with the abstraction process exclusively on the faces thereof are displayed, and a warning is also outputted.

5. A face identification method comprising the steps of:

detecting face images from human body images taken by a camera;

determining whether a detected face image matches with the face image previously stored by comparing both face images;

applying an abstraction process exclusively to a detected face image in order to make the detected face image unrecognizable, when it is determined that both face images do not match with each other; and

not applying the abstraction process to the detected face image when it is determined that both face images match with each other.