Image processing apparatus and method for detecting a feature point in an image

• US 7,903,880 B2
• Filed: 01/12/2007
• Issued: 03/08/2011
• Est. Priority Date: 01/13/2006
• Status: Expired due to Fees

• Current Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Original Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Inventors: Wyatt, Paul, Nakai, Hiroaki
• Primary Examiner(s): Werner; Brian P
• Application Number:
  US11/652,501
• Publication Number:
  US 20070183667A1
• Time in Patent Office: 1,516 Days
• Field of Search: None
• US Class Current: 382/199
• CPC Class Codes: G06T 7/13 Edge detection
• CPC Class Codes: G06V 10/443 by matching or filtering
• CPC Class Codes: G06V 10/46 Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features colour feature extraction G06V10/56

Abstract

An apparatus includes: a spatial differential value calculation unit that calculates a primary spatial differential value and a secondary spatial differential value of a brightness value with respect to each of a plurality of directions in each pixel of the image; a maximum and minimum direction estimation unit that estimates a maximum direction and a minimum direction for each pixel, the maximum direction being a direction with which the primary spatial differential value is a maximum value, the minimum direction being a direction with which the primary spatial differential value is a minimum value; and an evaluated value calculation unit that calculates a first evaluated value of each pixel using the primary spatial differential value in the maximum direction, the primary spatial differential value in the minimum direction, and the secondary spatial differential value in the maximum direction.

Claims

1. An image processing apparatus for calculating an evaluated value for detecting a feature point in an image, the apparatus comprising a processor that executes:

an image input unit for inputting an image;

a spatial differential value calculation unit that calculates a primary spatial differential value of a brightness value and a secondary spatial differential value of the brightness value with respect to each of a plurality of directions in each pixel of the image;

a maximum and minimum direction estimation unit that estimates a maximum direction and a minimum direction for each pixel, the maximum direction being a direction with which the primary spatial differential value is a maximum value, the minimum direction being a direction with which the primary spatial differential value is a minimum value;

an evaluated value calculation unit that calculates a first evaluated value of each pixel using the primary spatial differential value in the maximum direction in the pixel, the primary spatial differential value in the minimum direction in the pixel, and the secondary spatial differential value in the maximum direction in the pixel; and

an output unit that outputs the first evaluated value of each pixel.

2. The apparatus according to claim 1, wherein the evaluated value calculation unit comprises:

a noise degree estimation unit that estimates a noise amount in each pixel using the primary spatial differential value in the minimum direction;

a temporary evaluated value calculation unit that calculates a temporary evaluated value using the primary spatial differential direction in the maximum direction and the secondary spatial differential value in the minimum direction; and

a normalization unit that normalizes the temporary evaluated value using the temporary evaluated value and the noise amount so as to calculate the first evaluated value.

3. The apparatus according to claim 2, wherein the noise degree estimation unit estimates the noise amount in each pixel using an average value of absolute values of primary spatial differential values in minimum directions in a plurality of pixels located within a predetermined range including the pixel in the image.

4. The apparatus according to claim 2, wherein the noise degree estimation unit estimates the noise amount in each pixel using an average value of secondary spatial differential values in maximum directions in a plurality of pixels located within a predetermined range including the pixel in the image.

5. The apparatus according to claim 1, wherein the evaluated value calculation unit comprises:

a noise degree estimation unit that estimates a noise amount in each pixel using the primary spatial differential value in the minimum direction; and

a temporary evaluated value calculation unit that calculates a first temporary evaluated value using the primary spatial differential direction in the maximum direction and the secondary spatial differential value in the minimum direction, the temporary evaluated value calculation unit calculating a second temporary evaluated value using a value obtained by subtracting the noise amount from the primary spatial differential value in the maximum direction and a value obtained by subtracting the noise amount from the secondary spatial differential value in the minimum direction, and

wherein the evaluated value calculation unit calculates the first evaluated value by dividing the second temporary evaluated value by the first temporary evaluated value.

6. The apparatus according to claim 5, wherein the noise degree estimation unit estimates the noise amount in each pixel using an average value of absolute values of primary spatial differential values in minimum directions in a plurality of pixels located within a predetermined range including the pixel in the image.

7. The apparatus according to claim 5, wherein the noise degree estimation unit estimates the noise amount in each pixel using an average value of secondary spatial differential values in maximum directions in a plurality of pixels located within a predetermined range including the pixel in the image.

8. The apparatus according to claim 1, wherein the maximum and minimum direction estimation unit estimates the maximum direction using the primary spatial differential values in two or more different directions, and determines a direction perpendicular to the maximum direction to be the minimum direction.

9. The apparatus according to claim 1, wherein the maximum and minimum direction estimation unit estimates the minimum direction using the primary spatial differential values in two or more different directions, and determines a direction perpendicular to the minimum direction to be the maximum direction.

10. The apparatus according to claim 1, wherein the maximum and minimum direction estimation unit estimates a secondary differential maximum direction and a secondary differential minimum direction in each pixel, the secondary differential maximum direction being a direction with which the secondary spatial differential value is a maximum value, a secondary differential minimum direction being a direction with which the secondary spatial differential value is a minimum value,

wherein the evaluated value calculation unit calculates a second evaluated value of each pixel using the secondary spatial differential value in the secondary differential maximum direction in the pixel, the secondary spatial differential value in the secondary differential minimum direction in the pixel, and the minimum value of the primary spatial differential value in the pixel, and

wherein the output unit also outputs the second evaluated value of each pixel.

11. An image processing method for calculating an evaluated value for detecting a feature point in an image, the method comprising:

inputting an image;

calculating, by a processor, a primary spatial differential value of a brightness value and a secondary spatial differential value of the brightness value with respect to each of a plurality of directions in each pixel of the image;

estimating, by the processor, a maximum direction and a minimum direction for each pixel, the maximum direction being a direction with which the primary spatial differential value is a maximum value, the minimum direction being a direction with which the primary spatial differential value is a minimum value;

calculating, by the processor, a first evaluated value of each pixel using the primary spatial differential value in the maximum direction in the pixel, the primary spatial differential value in the minimum direction in the pixel, and the secondary spatial differential value in the maximum direction in the pixel; and

outputting the first evaluated value of each pixel.

12. A non-transitory computer-readable medium that stores a computer program for causing a computer to perform a process for calculating an evaluated value for detecting a feature point in an image, the process comprising:

inputting an image;

calculating a primary spatial differential value of a brightness value and a secondary spatial differential value of the brightness value with respect to each of a plurality of directions in each pixel of the image;

estimating a maximum direction and a minimum direction for each pixel, the maximum direction being a direction with which the primary spatial differential value is a maximum value, the minimum direction being a direction with which the primary spatial differential value is a minimum value;

calculating a first evaluated value of each pixel using the primary spatial differential value in the maximum direction in the pixel, the primary spatial differential value in the minimum direction in the pixel, and the secondary spatial differential value in the maximum direction in the pixel; and

outputting the first evaluated value of each pixel.

13. An image processing apparatus for calculating an evaluated value for detecting a feature point in an image, the apparatus comprising a processor that executes:

an image input unit for inputting an image;

a spatial differential value calculation unit that calculates a primary spatial differential value of a brightness value and a secondary spatial differential value of the brightness value with respect to each of a plurality of directions in each pixel of the image;

a maximum and minimum direction estimation unit that estimates a maximum direction and a minimum direction for each pixel, the maximum direction being a direction with which the secondary spatial differential value is a maximum value, the minimum direction being a direction with which the secondary spatial differential value is a minimum value;

an evaluated value calculation unit that calculates an evaluated value of each pixel using the secondary spatial differential value in the maximum direction in the pixel, the secondary spatial differential value in the minimum direction in the pixel, and the primary spatial differential value in the maximum direction in the pixel; and

an output unit that outputs the evaluated value of each pixel.

14. The apparatus according to claim 13, wherein the evaluated value calculation unit comprises:

a noise degree estimation unit that estimates a noise amount in each pixel using the minimum value of the primary spatial differential value; and

a temporary evaluated value calculation unit that calculates a first temporary evaluated value using the secondary spatial differential value in the secondary differential maximum direction and the secondary spatial differential value in the minimum direction, and calculates a second temporary evaluated value using a value obtained by subtracting the noise amount from the secondary spatial differential value in the secondary differential maximum direction and a value obtained by subtracting the noise amount from the secondary spatial differential value in the minimum direction, and

wherein the evaluated value calculation unit calculates the evaluated value by dividing the second temporary evaluated value by the first temporary evaluated value.

15. An image processing method for calculating an evaluated value for detecting a feature point in an image, the method comprising:

inputting an image;

calculating, by a processor, a primary spatial differential value of a brightness value and a secondary spatial differential value of the brightness value with respect to each of a plurality of directions in each pixel of the image;

estimating, by the processor, a maximum direction and a minimum direction for each pixel, the maximum direction being a direction with which the secondary spatial differential value is a maximum value, the minimum direction being a direction with which the secondary spatial differential value is a minimum value;

calculating, by the processor, an evaluated value of each pixel using the secondary spatial differential value in the maximum direction in the pixel, the secondary spatial differential value in the minimum direction in the pixel, and the primary spatial differential value in the maximum direction in the pixel; and

outputting the evaluated value of each pixel.

16. A non-transitory computer-readable medium that stores a computer program for causing a computer to perform a process for calculating an evaluated value for detecting a feature point in an image, the process comprising:

inputting an image;

calculating a primary spatial differential value of a brightness value and a secondary spatial differential value of the brightness value with respect to each of a plurality of directions in each pixel of the image;

estimating a maximum direction and a minimum direction for each pixel, the maximum direction being a direction with which the secondary spatial differential value is a maximum value, the minimum direction being a direction with which the secondary spatial differential value is a minimum value;

calculating an evaluated value of each pixel using the secondary spatial differential value in the maximum direction in the pixel, the secondary spatial differential value in the minimum direction in the pixel, and the primary spatial differential value in the maximum direction in the pixel; and

outputting the evaluated value of each pixel.