Image processing device, image processing method, and program

1. An image processing device comprising:

one or more processors operable to:

detect feature points from an image;

compute reliability of each of the detected feature points; and

remove at least one detected feature point from each split region of a plurality of split regions on the basis of the reliability so that a number of remaining feature points in each of the plurality of split regions is within a predetermined number associated with corresponding split region being processed, wherein the predetermined number for each of the split regions is determined based on position of the corresponding split region within the image, wherein the plurality of split regions is obtained by splitting the image, wherein the predetermined number for each of the split regions is determined based on position of an object within the image, wherein a first split region located within a specific range of the object has a predetermined number larger than a second split region located at a position outside the specific range.

2. The image processing device according to claim 1 , wherein the one or more processors are operable to remove the detected feature points sequentially from each of the plurality of split regions in order of increasing reliability.

3. The image processing device according to claim 1 , wherein the one or more processors are operable to set the predetermined number for each of plurality of the split regions.

4. The image processing device according to claim 1 , wherein

a feature point is a corner from the image, and

the reliability is computed based on a strength of the corner.

5. The image processing device according to claim 1 , wherein the one or more processors are operable to:

detect the feature points based on scale images, wherein the scale images are generated by smoothing the image with a predetermined smoothness in a stepwise manner; and

compute the reliability based on a difference value between the scale images.

6. The image processing device according to claim 1 , wherein the one or more processors are operable to split a three-dimensional region into the plurality of split regions.

7. The image processing device according to claim 6 , wherein the three-dimensional region comprises different scale two-dimensional images.

8. An image processing method comprising:

detecting feature points from an image;

computing reliability of each of the detected feature points; and

removing at least one detected feature point from each split region of a plurality of split regions on the basis of the reliability so that a number of remaining feature points in each of the plurality of split regions is within a predetermined number associated with corresponding split region being processed, wherein the predetermined number for each of the split regions is determined based on position of the corresponding split region within the image, wherein the plurality of split regions is obtained by splitting the image, wherein the predetermined number for each of the split regions is determined based on position of an object within the image, wherein a first split region located within a specific range of the object has the predetermined number larger than a second split region located at a position outside the specific range.

9. A non-transitory computer-readable storage medium having stored thereon, a computer program having at least one code section for image processing, the at least one code section being executable by a device for causing the device to perform steps comprising:

in an image processing device:

detecting feature points from an image;

computing reliability of each of the detected feature points; and

removing at least one detected feature point from each split region of a plurality of split regions on the basis of the reliability so that a number of remaining feature points in each of the plurality of split regions is within a predetermined number associated with corresponding split region being processed, wherein the predetermined number for each of the split regions is determined based on position of the corresponding split region within the image, wherein the plurality of split regions is obtained by splitting the image, wherein the predetermined number for each of the split regions is determined based on position of an object within the image, wherein a first split region located within a specific range of the object has the predetermined number larger than a second split region located at a position outside the specific range.

10. The non-transitory computer-readable storage medium according to claim 9 , wherein the detected feature points are sequentially removed from each of the split regions of the plurality of split regions in order of increasing reliability.

11. The non-transitory computer-readable storage medium according to claim 9 , wherein

a feature point is a corner from the image, and

the reliability is computed based on a strength of the corner.

12. The non-transitory computer-readable storage medium according to claim 9 , wherein

the feature points are detected based on scale images, wherein the scale images are generated by smoothing the image with a predetermined smoothness in a stepwise manner, and

the reliability is computed based on a difference value between the scale images.

13. The image processing method according to claim 8 , further comprising sequentially removing the detected feature points from each of the plurality of split regions in order of increasing reliability.

14. The image processing method according to claim 8 , wherein

a feature point is a corner from the image, and

the reliability is computed based on a strength of the corner.

15. The image processing method according to claim 8 , further comprising:

detecting the feature points based on scale images, wherein the scale images are generated by smoothing the image with a predetermined smoothness in a stepwise manner; and

computing the reliability based on a difference value between the scale images.