Apparatus and method for detecting key point using high-order laplacian of gaussian (LoG) kernel

• US 9,842,273 B2
• Filed: 03/09/2016
• Issued: 12/12/2017
• Est. Priority Date: 03/09/2015
• Status: Expired due to Fees

• Current Assignee: Electronics and Telecommunications Research Institute
• Original Assignee: Electronics and Telecommunications Research Institute
• Inventors: Cho, Yong Ju, Park, Unsang, Seok, Joo Myoung, Ahn, Sang Woo, Cha, Ji Hun
• Primary Examiner(s): PARK, CHAN S
• Assistant Examiner(s): Thomas, Mia M
• Application Number:
  US15/064,800
• Publication Number:
  US 20160267347A1
• Time in Patent Office: 643 Days
• CPC Class Codes: G06V 10/462 Salient features, e.g. scale invariant feature transforms [SIFT]

Abstract

Provided is an apparatus and method for detecting a key point using a high-order Laplacian of Gaussian (LoG) kernel. The high-order LoG kernel is generated based on an LoG operator which is calculated by sequentially differentiating an LoG operator with respect to x and y of an image. A scale space is generated based on the high-order LoG kernel and the key point is detected by comparing a current pixel in the scale space to pixels adjacent to the current pixel.

Claims

1. A method of detecting a key point of an image, the method comprising:

calculating a high-order Laplacian of Gaussian (LoG) operator by sequentially differentiating an LoG operator with respect to x and v of an image;

generating the high-order LoG kernel based on the high-order LoG operator;

generating a scale space based on the high-order LoG kernel; and

detecting the key point by comparing a current pixel in the scale space to pixels adjacent to the current pixel,

wherein the high-order LoG kernel is a mask for performing a convolution on a function using the high-order LoG operator.

2. The method of claim 1, wherein the high-order LoG operator is a high-order differential coefficient of the LoG operator.

4. The method of claim 1, further comprising:

searching an image corresponding to a current image based on the key point.

5. The method of claim 4, wherein the searching of the image comprises searching for a number of key points of which a similarity between a key point detected from a current image and a key point detected from another image is greater than or equal to a threshold value, and determining another image of which the found number of key points is greater than or equal to the threshold value, as the image corresponding to the current image.

3. The method of 1, wherein the generating of the scale space comprises generating the scale space by performing a convolution operation on the image by changing a σ

value comprised in a high-order LoG operator to kσ

, k²σ

, k³σ

, and k⁴σ

, using a predetermined constant k.

6. An apparatus for detecting a key point of an image, the apparatus comprising:

a processor configured to generate a high-order Laplacian of Gaussian (LoG) kernel using an LoG operator and detect a key point of an image using the high-order LoG kernel,

wherein the processor is configured to calculate a high-order LoG operator by sequentially differentiating the LoG operator with respect to x and y of an image and generate the high-order LoG kernel based on the high-order LoG operator and generate a scale space based on the high-order LoG kernel and detect the key point by comparing a current pixel in the scale space to pixels adjacent to the current pixel, and

the high-order LoG kernel is a mask for performing a convolution on a function using the high-order LoG operator.

7. The apparatus of claim 6, wherein the high-order LoG operator is a high-order differential coefficient of the LoG operator.

8. The apparatus of claim 6, wherein the processor is configured to generate the scale space by performing a convolution operation on the image by changing a σ

value comprised in a high-order LoG operator to kσ

, k²σ

, k³σ

, and k⁴σ

, using a predetermined constant k.

9. The apparatus of claim 6, wherein the processor is configured to search an image corresponding to a current image based on the key point.

10. The apparatus of claim 9, wherein the processor is configured to search for a number of key points of which a similarity between a key point detected from a current image and a key point detected from another image is greater than or equal to a threshold value, and determine another image of which the found number of key points is greater than or equal to the threshold value, as the image corresponding to the current image.