IMAGE ATTRIBUTE DISCRIMINATION APPARATUS, ATTRIBUTE DISCRIMINATION SUPPORT APPARATUS, IMAGE ATTRIBUTE DISCRIMINATION METHOD, ATTRIBUTE DISCRIMINATION SUPPORT APPARATUS CONTROLLING METHOD, AND CONTROL PROGRAM

• US 20110222775A1
• Filed: 02/23/2011
• Published: 09/15/2011
• Est. Priority Date: 03/15/2010
• Status: Active Grant

• Current Assignee: Omron Corporation
• Original Assignee: Omron Corporation
• Inventors: Ruan, Xiang, Tanaka, Kiyoaki, Yamamoto, Motoo, Aizawa, Tomoyoshi
• Granted Patent: US 9,177,205 B2
• Time in Patent Office: Days
• US Class Current: 382/190
• CPC Class Codes: G06V 20/10 Terrestrial scenes scenes under surveillance with static cameras G06V20/52; scenes perceived from the exterior of a vehicle G06V20/56; scenes perceived from the interior of a vehicle G06V20/59
• CPC Class Codes: G06V 20/635 Overlay text, e.g. embedded captions in a TV program

Abstract

An attribute of image data can accurately be discriminated. An image attribute discrimination apparatus includes a heterogeneous region extracting unit that specifies a heterogeneous region from image data. The heterogeneous region includes a heterogeneous matter whose attribute is different from that of a content originally produced by the image data. An image attribute discrimination apparatus further includes a scene discrimination unit that discriminates the attribute of the image data based on a feature quantity extracted from a pixel group except each pixel in the heterogeneous region in each pixel of the image data.

Claims

1. An image attribute discrimination apparatus that discriminates an attribute of image data based on a content produced by the image data, the image attribute discrimination apparatus comprising:

a heterogeneous region specifying unit for specifying a heterogeneous region from the image data, wherein the heterogeneous region comprises a heterogeneous matter whose attribute is different from that of the content originally produced by the image data; and

an attribute discrimination unit for discriminating the attribute of the image data based on a feature quantity extracted from a pixel group except each pixel in the heterogeneous region in each pixel of the image data.

2. The image attribute discrimination apparatus according to claim 1, wherein the heterogeneous region specifying unit specifies a character region including a character as the heterogeneous region.

3. The image attribute discrimination apparatus according to claim 2, further comprising:

a restoration unit for restoring a pixel identical to a target pixel with respect to an out-of-target pixel region in each pixel of the image data, the out-of-target region that does not become a feature quantity extracting target.

4. The image attribute discrimination apparatus according to claim 3, further comprising a target pixel determination unit for determining whether each pixel in the heterogeneous region specified by the heterogeneous region specifying unit is the target pixel whose feature quantity is extracted by the attribute discrimination unit,

wherein the attribute discrimination unit discriminates the attribute of the image data based on the feature quantity extracted from the pixel group except the out-of-target pixel in each pixel of the image data, wherein the out-of-target pixel being determined to be out of the target by the target pixel determination unit.

5. The image attribute discrimination apparatus according to claim 4,

wherein the heterogeneous region specifying unit specifies the character region including the character as the heterogeneous region,

wherein the image attribute discrimination apparatus further includes a character recognition unit for recognizing the character in the character region specified by the heterogeneous region specifying unit, and

wherein the target pixel determination unit determines the pixel in the character region as the target pixel, when a degree of reliability of a character recognition result is not more than a predetermined value, wherein the degree of reliability indicates likelihood that the character in the character region is the character recognized by the character recognition unit.

6. The image attribute discrimination apparatus according to claim 5, further comprising:

a keyword extracting unit for extracting a keyword, the character or character string recognized by the character recognition unit; and

a word association storage unit in which association between each keyword extracted by the keyword extracting unit and each attribute discriminated by the attribute discrimination unit is stored,

wherein the attribute discrimination unit refers to the word association storage unit, and

wherein the attribute discrimination unit discriminates the attribute of the image data in consideration of a level of association between the keyword extracted from the character region of the image data and each attribute.

7. The image attribute discrimination apparatus according to claim 6,

wherein the attribute discrimination unit checks the feature quantity of the image data against a model feature quantity that is previously defined in each plurality of kinds of attributes,

wherein the attribute discrimination unit discriminates the attribute of the image data by computing a degree of reliability of an attribute discrimination result according to a degree of similarity between the feature quantity of the image data and the model feature quantity, the degree of reliability indicating likelihood that the attribute of the image data is the attribute, and

wherein the association between the keyword and the attribute is stored in the word association storage unit as a score added to the degree of reliability of the attribute discrimination result.

8. The image attribute discrimination apparatus according to claim 7,

wherein the restoration unit performs the restoration when the degree of reliability of the attribute discrimination result is lower than a predetermined value.

9. The image attribute discrimination apparatus according to claim 7, wherein the attribute discrimination unit computes the degree of reliability lower with increasing region of the out-of-target pixel that does not becomes the feature quantity extracting target in each pixel of the image data.

10. The image attribute discrimination apparatus according to claim 4, wherein the target pixel determination unit determines each pixel in the heterogeneous region as the out-of-target pixel only when an area occupied by the heterogeneous region in the image data is more than a predetermined value.

11. The image attribute discrimination apparatus according to claim 1, further comprising:

a model feature quantity computing unit for computing a model feature quantity of a designated attribute using the feature quantity extracted from the pixel group except each pixel in the heterogeneous region specified by the heterogeneous region specifying unit in each pixel of the image data, when image data and the designation of the attribute of the image data are inputted to the image attribute discrimination apparatus,

wherein the attribute discrimination unit checks the feature quantity of the image data against the model feature quantity computed in each attribute by the model feature quantity computing unit, and

wherein the attribute discrimination unit discriminates the attribute of the image data according to a degree of similarity between the feature quantity of the image data and the model feature quantity.

12. An attribute discrimination support apparatus that defines a model feature quantity in each attribute, an image attribute discrimination apparatus referring to the model feature quantity, the image attribute discrimination apparatus discriminating an attribute of image data based on a content produced by the image data, the attribute discrimination support apparatus comprising:

a heterogeneous region specifying unit for specifying a heterogeneous region from image data when the image data and designation of the attribute of the image data are inputted, wherein the heterogeneous region further comprises a heterogeneous matter whose attribute is different from that of the content originally produced by the image data; and

a model feature quantity computing unit for computing a model feature quantity of the designated attribute using a feature quantity from a pixel group except each pixel in the heterogeneous region in each pixel of the image data.

13. An image attribute discrimination method for discriminating an attribute of image data based on a content produced by the image data, the image attribute discrimination method comprising the steps of:

specifying a heterogeneous region from the image data, wherein the heterogeneous region further comprises a heterogeneous matter whose attribute is different from that of the content originally produced by the image data; and

discriminating the attribute of the image data based on a feature quantity extracted from a pixel group except each pixel in the heterogeneous region in each pixel of the image data.

15. A control program that causes a computer to perform the steps of the image attribute discrimination method according to claim 13.

14. A method for controlling an attribute discrimination support apparatus that defines a model feature quantity in each attribute, an image attribute discrimination apparatus referring to the model feature quantity, the image attribute discrimination apparatus discriminating an attribute of image data based on a content produced by the image data, the discrimination support apparatus controlling method comprising the steps of:

specifying a heterogeneous region from image data when the image data and designation of the attribute of the image data are inputted, wherein the heterogeneous region further comprises a heterogeneous matter whose attribute is different from that of the content originally produced by the image data; and

computing a model feature quantity of the designated attribute using a feature quantity from a pixel group except each pixel in the heterogeneous region in each pixel of the image data.

16. A control program that causes a computer to perform the steps of the attribute discrimination support apparatus controlling method according to claim 14.