Generating feature descriptors for image analysis

1. A descriptor generation system for generating a rotation-invariant feature descriptor for a location in an image for use in performing descriptor matching in analysing the image, the descriptor generation system being configured to:

extract samples according to a descriptor pattern for the location in the image;

use the extracted samples to determine a measure of rotation for the location in the image, the measure of rotation describing an angle between an orientation of the image and a characteristic direction of the image at the location;

the descriptor generation system further comprising a feature descriptor generator configured to:

generate a feature descriptor for the location in the image by determining a set of samples characterising the location in dependence on the determined measure of rotation and the extracted samples.

2. A descriptor generation system according to claim 1 , in which the descriptor generation system is further configured to determine the set of samples by sampling from the extracted samples.

3. A descriptor generation system according to claim 1 , in which the descriptor pattern comprises at least one ring surrounding the location in the image.

4. A descriptor generation system according to claim 3 , in which the at least one ring is one of a circle, a wavy circle and a polygon.

5. A descriptor generation system according to claim 1 , in which the descriptor generation system is configured to store the extracted samples in an array.

6. A descriptor generation system according to claim 5 , in which the array is a linear array.

7. A descriptor generation system according to claim 5 , in which the descriptor generation system is configured to determine the set of samples by shifting elements of at least one portion of the array along a number of positions in the respective portion of the array, the number of positions being determined in dependence on the determined measure of rotation.

8. A descriptor generation system according to claim 1 , in which the descriptor generation system is configured to determine the set of samples by interpolating between two or more samples of the extracted samples.

9. A descriptor generation system according to claim 8 , in which the samples of the extracted samples between which interpolation is performed have been obtained from one or more ring of the descriptor pattern.

10. A descriptor generation system according to claim 9 , in which the descriptor generation system is configured to determine the set of samples by interpolating between the samples of the extracted samples along a portion of the shape of the ring to which the samples correspond.

11. A descriptor generation system according to claim 9 , in which the samples of the extracted samples between which interpolation is performed have been obtained from adjacent rings in the descriptor pattern.

12. A descriptor generation system according to claim 9 , in which the samples of the extracted samples comprise N 1 samples from a first ring and N 2 samples from a second ring, where N 1 ≤N 2 , in which the first ring is radially inside the second ring.

13. A descriptor generation system according to claim 1 , in which the descriptor generation system is configured to determine the set of samples by at least one of:

a linear interpolation between samples of the extracted samples;

a quadratic interpolation between samples of the extracted samples; and

a cubic interpolation between samples of the extracted samples.

14. A descriptor generation system according to claim 1 , in which the location in the image is one or more of:

a point in the image; and

a pixel location in the image.

15. A descriptor generation system according to claim 1 , in which the descriptor generation system is configured to discard the set of samples once the feature descriptor has been generated.

16. A descriptor generation system according to claim 1 , in which the descriptor generation system is configured to extract samples by sampling scale-space data representative of the image at a plurality of length scales.

17. A descriptor generation system according to claim 1 , in which the descriptor generation system is configured to extract samples by filtering at a sampling point using a filter sized to overlap with a filter at a neighbouring sampling point.

18. A descriptor generation system according to claim 1 , in which the descriptor generation system is configured to:

store the extracted samples in an array and to determine the set of samples by shifting elements of at least one portion of the array along a number of positions in the respective portion of the array, the number of positions being determined in dependence on the determined measure of rotation; and

determine the set of samples by interpolating between two or more samples of the extracted samples.

19. A computer-implemented method for generating a rotation-invariant feature descriptor for a location in an image for use in performing descriptor matching in analysing the image, the method comprising:

extracting samples according to a descriptor pattern for the location in the image;

using the extracted samples to determine a measure of rotation for the location in the image, the measure of rotation describing an angle between an orientation of the image and a characteristic direction of the image at the location;

generating a feature descriptor for the location in the image by determining a set of samples characterising the location in dependence on the determined measure of rotation and the extracted samples; and

processing the determined set of samples to generate the feature descriptor for the location in the image.

20. A non-transitory computer readable storage medium having stored thereon computer readable instructions that, when executed at a computer system, cause the computer system to:

extract samples according to a descriptor pattern for the location in the image;

use the extracted samples to determine a measure of rotation for the location in the image, the measure of rotation describing an angle between an orientation of the image and a characteristic direction of the image at the location;

generate a feature descriptor for the location in the image by determining a set of samples characterising the location in dependence on the determined measure of rotation and the extracted samples; and

process the determined set of samples to generate the feature descriptor for the location in the image.