IMAGE RECOGNITION METHOD AND IMAGE RECOGNITION APPARATUS

• US 20120076417A1
• Filed: 06/15/2010
• Published: 03/29/2012
• Est. Priority Date: 06/18/2009
• Status: Active Grant

• Current Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Original Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Inventors: Yoshii, Hiroto, Matsugu, Masakazu
• Granted Patent: US 9,852,159 B2
• Time in Patent Office: Days
• US Class Current: 382/190
• CPC Class Codes: G06F 16/51 Indexing; Data structures therefor; Storage structures
• CPC Class Codes: G06F 18/24323 Tree-organised classifiers
• CPC Class Codes: G06F 18/254 of classification results, e.g. of results related to same input data
• CPC Class Codes: G06F 18/28 Determining representative reference patterns, e.g. by averaging or distorting; Generating dictionaries
• CPC Class Codes: G06T 2207/20021 Dividing image into blocks, subimages or windows
• CPC Class Codes: G06T 2207/20081 Training; Learning
• CPC Class Codes: G06T 7/337 involving reference images or patches
• CPC Class Codes: G06V 10/50 by performing operations within image blocks; by using histograms, e.g. histogram of oriented gradients [HoG]; by summing image-intensity values; Projection analysis
• CPC Class Codes: G06V 10/751 Comparing pixel values or logical combinations thereof, or feature values having positional relevance, e.g. template matching
• CPC Class Codes: G06V 10/764 using classification, e.g. of video objects
• CPC Class Codes: G06V 10/772 Determining representative reference patterns, e.g. averaging or distorting patterns; Generating dictionaries
• CPC Class Codes: G06V 10/809 of classification results, e.g. where the classifiers operate on the same input data

Abstract

An image recognition apparatus is provided which comprises a first extracting means for extracting, from every registration image previously registered, a set of registration partial images of a predetermined size, and a second extracting means for extracting, from an input new image, a set of new partial images of a predetermined size. The apparatus further comprises a discriminating means for discriminating an attribute of the new partial image based on a rule formed by dividing the set of the registration partial images extracted by the first extracting means, and a collecting means for deriving a final recognition result of the new image by collecting discrimination results by the discriminating means at the time when the new partial images as elements of the set of the new partial images are input.

Claims

1. An image recognition method comprising:

a first extracting step of extracting, from every registration image which is previously registered, a set of registration partial images of a predetermined size;

a second extracting step of extracting, from an input new image, a set of new partial images of a predetermined size;

a discriminating step of discriminating an attribute of the new partial image based on a rule which is formed by dividing the set of the registration partial images extracted in the first extracting step; and

a collecting step of deriving a final recognition result of the new image by collecting discrimination results in the discriminating step at the time when the new partial images as elements of the set of the new partial images are input.

2. An image recognition method according to claim 1, wherein the attribute is a probability in which the new partial image is included in each of the registration images.

3. An image recognition method according to claim 1, wherein the attribute is a probability in which the new partial image is included in each of the registration partial images.

4. An image recognition method according to claim 1, wherein in the discriminating step, a plurality of rules are used, a result of the discrimination made based on each of the rules is collected, and a final attribute is decided.

5. An image recognition method according to claim 1, wherein in the discriminating step, the division of the set is performed based on a luminance comparison of two pixels selected at random.

6. An image recognition method according to claim 1, wherein the rule is formed by recursively dividing the set of the registration partial images.

7. An image recognition method according to claim 1, further comprising a deciding step of deciding the predetermined partial image size by using a cross validation based on the image recognition method.

8. An image recognition method according to claim 1, further comprising a deciding step of deciding the predetermined partial image size based on a value obtained by dividing an average of distances between an average partial image obtained by averaging the registration partial images in each registration image and the respective registration partial images by a distance between average partial images of the different registration images.

9. An image recognition method comprising:

an extracting step of extracting a partial image from an input image;

a discriminating step of discriminating to which partial image of which category that has previously been learned the partial image belongs;

a voting step of voting based on a discrimination result in the discriminating step;

a collecting step of collecting voting results obtained by repeating the extracting step, the discriminating step, and the voting step; and

an identifying step of identifying a position and a category of an object existing in the input image based on a collection result in the collecting step.

10. An image recognition method according to claim 9, wherein the category is an orientation of the object.

11. An image recognition method according to claim 10, wherein in the voting step, a vote is performed to an orientation discriminated in the discriminating step and to an orientation similar to the orientation.

12. An image recognition method according to claim 9, wherein the discriminating step is executed by using one or a plurality of classification trees.

13. An image recognition method according to claim 12, wherein the plurality of classification trees are used to discriminate a same category.

14. An image recognition method according to claim 12, wherein the plurality of classification trees are used to discriminate different categories.

15. An image recognition apparatus comprising:

a first extracting unit configured to extract, from every registration image which is previously registered, a set of registration partial images of a predetermined size;

a second extracting unit configured to extract, from an input new image, a set of new partial images of a predetermined size;

a discriminating unit configured to discriminate an attribute of the new partial image based on a rule which is formed by dividing the set of the registration partial images extracted by the first extracting unit; and

a collecting unit configured to derive a final recognition result of the new image by collecting discrimination results by the discriminating unit at the time when the new partial images as elements of the set of the new partial images are input.

16. An image recognition apparatus comprising:

an extracting unit configured to extract a partial image from an input image;

a discriminating unit configured to discriminate to which partial image of which category that has previously been learned the partial image belongs;

a voting unit configured to vote based on a discrimination result by the discriminating unit;

a collecting unit configured to collect voting results obtained by repeating processing by the extracting unit, the discriminating unit, and the voting unit; and

an identifying unit configured to identify a position and a category of an object existing in the input image based on a collection result by the collecting unit.

17. A non-transitory computer-readable storage medium storing a computer program for causing a computer to execute:

a first extracting step of extracting, from every registration image which is previously registered, a set of registration partial images of a predetermined size;

a second extracting step of extracting, from an input new image, a set of new partial images of a predetermined size;

a discriminating step of discriminating an attribute of the new partial image based on a rule which is formed by dividing the set of the registration partial images extracted in the first extracting step; and

a collecting step of deriving a final recognition result of the new image by collecting discrimination results in the discriminating step at the time when the new partial images as elements of the set of the new partial images are input.

18. A non-transitory computer-readable storage medium storing a computer program for causing a computer to execute:

an extracting step of extracting a partial image from an input image;

a discriminating step of discriminating to which partial image of which category that has previously been learned the partial image belongs;

a voting step of voting based on a discrimination result in the discriminating step;

a collecting step of collecting voting results obtained by repeating the extracting step, the discriminating step, and the voting step; and

an identifying step of identifying a position and a category of an object existing in the input image based on a collection result in the collecting step.