IMAGE ANALYZING DEVICE, IMAGE ANALYZING METHOD, AND RECORDING MEDIUM STORING IMAGE ANALYZING PROGRAM

• US 20150146924A1
• Filed: 11/14/2014
• Published: 05/28/2015
• Est. Priority Date: 11/27/2013
• Status: Active Grant

• Current Assignee: Ricoh Company Limited
• Original Assignee: Hiroshi Shimura, Yuka Kihara
• Inventors: KIHARA, Yuka, Shimura, Hiroshi
• Granted Patent: US 9,613,272 B2
• Time in Patent Office: Days
• US Class Current: 382/103
• CPC Class Codes: G06F 18/2413 based on distances to training or reference patterns
• CPC Class Codes: G06F 18/2415 based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
• CPC Class Codes: G06V 10/255 Detecting or recognising potential candidate objects based on visual cues, e.g. shapes
• CPC Class Codes: G06V 10/764 using classification, e.g. of video objects
• CPC Class Codes: G06V 20/00 Scenes; Scene-specific elements control of digital cameras H04N23/60

Abstract

An image analyzing device, an image analyzing method, and a recording medium storing an image analyzing program are provided. Each of the image analyzing device, the image analyzing method, and the recording medium storing the image analyzing program recognizes an area where a target is displayed based on a feature value of an input image, generates space recognition information to recognize spatial properties of each portion of the input image, divides the image into a plurality of similar areas according to similarity in feature value of the input image, the similar area having a similar feature value, obtains specified attribute data of the spatial properties to be referred to, from image areas around the recognized area where the target is displayed, recognizes the spatial properties according to the space recognition information, and determines whether a result of recognition is appropriate at the portion where the target is displayed.

Claims

1. An image analyzing device comprising:

a target recognition unit to recognize an area where a target is displayed based on a feature value of an input image to generate a recognition result;

a space recognition unit to generate space recognition information to recognize spatial properties of each portion of the input image;

a feature dividing unit to divide the image into a plurality of similar areas according to similarity in feature value of the input image, the similar area having a similar feature value;

a space specification data obtainer to obtain specified attribute data of the spatial properties to be referred to, from image areas around the recognized area where the target is displayed; and

a recognition finalizing unit to recognize the spatial properties according to the space recognition information, and determine whether the recognition result is appropriate at the portion where the target is displayed, based on the distribution of the similar areas of the specified spatial properties in the image areas around the recognized area where the target is displayed.

2. The image analyzing device according to claim 1, further comprising:

a learning information storage unit to store learning information including the distribution of the similar areas around the area where the target is displayed and the specified attribute data of the spatial properties to be referred to,

wherein

the distribution of the similar areas in the learning information is on an area of the specified spatial properties, and

the recognition finalizing unit determines whether the recognition result is appropriate at the portion where the target is displayed, based on an input image distribution indicating the distribution of the similar areas around the recognized area where the target is displayed on an input image, and based on a distribution of learnt image indicating the distribution of the similar areas in the learning information.

3. The image analyzing device according to claim 1, wherein the recognition finalizing unit performs calculation in such a manner that a greater value is obtained as the similar area being distributed on the distribution of learnt image is distributed on the input image distribution to a greater degree, and determines that the recognition result is appropriate at the portion where the target is displayed when the obtained value is greater than a prescribed threshold.

4. The image analyzing device according to claim 3, wherein the recognition finalizing unit performs calculation in such a manner that a greater value is obtained as the similar area being distributed to a small degree on the distribution of the similar areas on an entirety of the input image is distributed to a greater degree on the input image distribution.

5. The image analyzing device according to claim 1, wherein the space recognition unit recognizes one attribute of an under side of space, an upper side of the space, and an upright object disposed between the under side and the upper side.

6. A method of analyzing an image, the method comprising:

recognizing an area where a target is displayed based on a feature value of an input image to generate a recognition result;

generating space recognition information to recognize spatial properties of each portion of the input image;

dividing the image into a plurality of similar areas according to similarity in feature value of the input image, the similar area having a similar feature value;

obtaining specified attribute data of the spatial properties to be referred to, from image areas around the recognized area where the target is displayed;

recognizing the spatial properties according to the space recognition information; and

determining whether the recognition result is appropriate at the portion where the target is displayed, based on the distribution of the similar areas of the specified spatial properties in the image areas around the recognized area where the target is displayed.

7. The method according to claim 6, further comprising:

storing learning information including the distribution of the similar areas around the area where the target is displayed and the specified attribute data of the spatial properties to be referred to,

wherein

the distribution of the similar areas in the learning information is on an area of the specified spatial properties, and

the determining includes determining whether the recognition result is appropriate at the portion where the target is displayed, based on an input image distribution indicating the distribution of the similar areas around the recognized area where the target is displayed on an input image, and based on a distribution of learnt image indicating the distribution of the similar areas in the learning information.

8. The method according to claim 7, wherein the determining includes performing calculation in such a manner that a greater value is obtained as the similar area being distributed on the distribution of learnt image is distributed on the input image distribution to a greater degree, and determining that the recognition result is appropriate at the portion where the target is displayed when the obtained value is greater than a prescribed threshold.

9. The method according to claim 8, wherein the determining includes performing calculation in such a manner that a greater value is obtained as the similar area being distributed to a small degree on the distribution of the similar areas on an entirety of the input image is distributed to a greater degree on the input image distribution.

10. The method according to claim 6, wherein the generating includes recognizing one attribute of an under side of space, an upper side of the space, and an upright object disposed between the under side and the upper side.

11. A computer-readable non-transitory recording medium storing a program for causing a computer to execute a method of analyzing an image, the method comprising:

recognizing an area where a target is displayed based on a feature value of an input image to generate a recognition result;

generating space recognition information to recognize spatial properties of each portion of the input image;

dividing the image into a plurality of similar areas according to similarity in feature value of the input image, the similar area having a similar feature value;

obtaining specified attribute data of the spatial properties to be referred to, from image areas around the recognized area where the target is displayed;

recognizing the spatial properties according to the space recognition information; and

determining whether the recognition result is appropriate at the portion where the target is displayed, based on the distribution of the similar areas of the specified spatial properties in the image areas around the recognized area where the target is displayed.

12. The method according to claim 11, further comprising:

storing learning information including the distribution of the similar areas around the area where the target is displayed and the specified attribute data of the spatial properties to be referred to,

wherein

the distribution of the similar areas in the learning information is on an area of the specified spatial properties, and

the determining includes determining whether the recognition result is appropriate at the portion where the target is displayed, based on an input image distribution indicating the distribution of the similar areas around the recognized area where the target is displayed on an input image, and based on a distribution of learnt image indicating the distribution of the similar areas in the learning information.

13. The method according to claim 12, wherein the determining includes performing calculation in such a manner that a greater value is obtained as the similar area being distributed on the distribution of learnt image is distributed on the input image distribution to a greater degree, and determines that the recognition result is appropriate at the portion where the target is displayed when the obtained value is greater than a prescribed threshold.

14. The method according to claim 13, wherein the determining includes performing calculation in such a manner that a greater value is obtained as the similar area being distributed to a small degree on the distribution of the similar areas on an entirety of the input image is distributed to a greater degree on the input image distribution.

15. The method according to claim 11, wherein the generating includes recognizing one attribute of an under side of space, an upper side of the space, and an upright object disposed between the under side and the upper side.