Image filtering method for detecting orientation component of edge and image recognizing method using the same

• US 9,355,335 B2
• Filed: 02/20/2012
• Issued: 05/31/2016
• Est. Priority Date: 07/18/2011
• Status: Active Grant

• Current Assignee: COGA Robotics Company Limited
• Original Assignee: Industry-University Cooperation Foundation Hanyang University
• Inventors: Suh, Il-Hong, Park, Young-Bin
• Primary Examiner(s): Couso, Yon
• Application Number:
  US14/233,181
• Publication Number:
  US 20140153781A1
• Time in Patent Office: 1,562 Days
• Field of Search: 382/103, 382/199
• US Class Current: 1/1
• CPC Class Codes: G06T 2207/10004 Still image; Photographic image
• CPC Class Codes: G06T 2207/20192 Edge enhancement; Edge preservation
• CPC Class Codes: G06T 5/20 using local operators
• CPC Class Codes: G06T 7/13 Edge detection
• CPC Class Codes: G06V 10/443 by matching or filtering
• CPC Class Codes: G06V 10/751 Comparing pixel values or logical combinations thereof, or feature values having positional relevance, e.g. template matching

Abstract

The present disclosure relates to an image filtering method for detecting an orientation component of an edge and an image recognizing method using the same. The image filtering method includes receiving an original image, generating a plurality of first images by filtering the original image with filters respectively generated along a plurality of channels, generating a second image by selecting a channel having a maximum value for each image unit, from the generated first images, and generating an output image whose edge is detected so as to maintain the consistency of channel by filtering the second image with filters respectively generated along the plurality of channels to generate a plurality of third images and comparing the channel of the second image with the channels of the third images.

Claims

1. An image filtering method, comprising:

receiving an original image;

generating a plurality of first images by filtering the original image with filters respectively generated along a plurality of channels;

generating a second image by selecting a channel having a maximum value for each image unit, from the generated first images wherein the image unit represents a basic unit in operation of an image; and

generating an output image whose edge is detected by filtering the second image with filters respectively generated along the plurality of channels to generate a plurality of third images and comparing the channel of the second image with the channels of the third images.

2. The image filtering method according to claim 1, wherein said generating of an output image comprises:

generating the plurality of third images by filtering the second image with filters respectively generated along the plurality of channels;

selecting a channel having a maximum value for each image unit, from the generated third images;

comparing the channel selected for each image unit from the third images with the channel of the second image; and

generating an output image by adjusting a value of the corresponding image unit according to the comparison result to maintain the consistency of the channel of the second image and the selected channel of the third images.

3. The image filtering method according to claim 1,

wherein when the channel of the second image and the selected channel of the third images are identical to each other as a result of the comparison, a value of the corresponding image unit is set as a value of the output image.

4. The image filtering method according to claim 1,

wherein when the channel of the second image and the selected channel of the third images are not identical to each other as a result of the comparison, a value of the corresponding image unit is decreased and set as a value of the output image.

5. The image filtering method according to claim 1,

wherein the consistency of channel has a higher value when the channel for generating the second image is identical to the channel selected for generating the third image.

6. The image filtering method according to claim 1, further comprising:

feeding back the output image as an input value in the operation of generating the second image.

7. The image filtering method according to claim 6,

wherein the feeding-back operation is repeated a predetermined number of times.

8. The image filtering method according to claim 1,

wherein the channel represents a preset specific angle, and

wherein the filter selectively detects an image of the set specific angle from the input image.

9. The image filtering method according to claim 1,

wherein the image is filtered by convoluting an input image and a Gaussian filter generated along the channel.

15. A non-transitory computer-readable recording medium on which a program for executing the method defined in the claim 1 is recorded.

16. The image filtering method according to claim 1, wherein the mage unit includes a pixel.

10. An image recognizing method, comprising:

receiving an original image;

detecting an edge of a partial image from the original image by means of a predetermined filtering process; and

recognizing an object present in the original image by using the detected edge and characteristic images which are classified and stored in advance,

wherein the predetermined filtering process includes;

generating a plurality of first images by filtering the original image with filters respectively generated along a plurality of channels;

generating a second image by selecting a channel having a maximum value for each image unit, from the generated first images wherein the image unit represents a basic unit in operation of an image; and

generating an output image whose edge is detected by filtering the second image with filters respectively generated along the plurality of channels to generate a plurality of third images and comparing the channel of the second image with the channels of the third images.

11. The image recognizing method according to claim 10, wherein said generating of an output image comprises:

generating the plurality of third images by filtering the second image with filters respectively generated along the plurality of channels;

selecting a channel having a maximum value for each image unit, from the generated third images;

comparing the channel selected for each image unit from the third images with the channel of the second image; and

generating an output image by adjusting a value of the corresponding image unit according to the comparison result to maintain the consistency of the channel of the second image and the selected channel of the third images.

12. The image recognizing method according to claim 10, wherein the predetermined filtering process further includes:

feeding back the output image as an input value in the operation of generating the second image.

13. The image recognizing method according to claim 10,

wherein the channel represents a preset specific angle, and

wherein the filter selectively detects an image of the set specific angle from the input image.

14. The image recognizing method according to claim 10,

wherein said recognizing of an object present in the original image recognizes an object corresponding to the characteristic images of the partial region which are classified and stored in advance, by using a histogram or hierarchical learning model generated from an orientation component of the detected edge.

17. The image recognizing method according to claim 10, wherein the image unit includes a pixel.

18. The image filtering method according to claim 10, wherein said generating the output image comprises:

when the channel of the second image is identical to the selected channel from the generated third images;

setting the corresponding basic image unit as a value of the output image; and

generating the output image with the value of the output image;

when the channel of the second image is not identical to the selected channel from the generated third images;

decreasing the value of the corresponding basic image unit;

setting the decreased value of the corresponding basic image unit as the value of the output image; and

generating the output image with the value of the output image.

19. An image filtering method, comprising:

receiving an original image;

generating a plurality of first images by filtering the original image with filters respectively generated along a plurality of channels;

generating a second image by selecting a channel having a maximum value for each image unit, from the generated first images wherein the image unit represents a basic unit in operation of an image;

generating a plurality of third images by filtering the second image with filters respectively generated along the plurality of channels;

selecting a channel having a maximum value for each basic image unit, from the generated third images;

comparing the channel selected for each basic image unit from the third images with the channel of the second image;

when the channel of the second image is identical to the selected channel from the generated third images;

setting the corresponding basic image unit as a value of the output image; and

generating the output image with the value of the output image;

when the channel of the second image is not identical to the selected channel from the generated third images;

decreasing the value of the corresponding basic image unit;

setting the decreased value of the corresponding basic image unit as the value of the output image; and

generating the output image with the value of the output image.