METHOD AND APPARATUS FOR TRACKING AND RECOGNITION WITH ROTATION INVARIANT FEATURE DESCRIPTORS

• US 20110286627A1
• Filed: 11/18/2010
• Published: 11/24/2011
• Est. Priority Date: 11/19/2009
• Status: Active Grant

• Current Assignee: Nokia Corporation, Stanford University
• Original Assignee: Nokia Corporation, Stanford University
• Inventors: Chandrasekhar, Vijay, Grzeszczuk, Radek, Takacs, Gabriel, Girod, Bernd
• Granted Patent: US 8,687,891 B2
• Time in Patent Office: Days
• US Class Current: 382/103
• CPC Class Codes: G06T 2207/10016 Video; Image sequence
• CPC Class Codes: G06T 2207/20164 Salient point detection; Corner detection
• CPC Class Codes: G06T 7/246 using feature-based methods, e.g. the tracking of corners or segments
• CPC Class Codes: G06V 10/462 Salient features, e.g. scale invariant feature transforms [SIFT]
• CPC Class Codes: G06V 10/50 by performing operations within image blocks; by using histograms, e.g. histogram of oriented gradients [HoG]; by summing image-intensity values; Projection analysis
• CPC Class Codes: G06V 10/52 Scale-space analysis, e.g. wavelet analysis multi-scale boundary representations G06V10/42

Abstract

Various methods for tracking and recognition with rotation invariant feature descriptors are provided. One example method includes generating an image pyramid of an image frame, detecting a plurality of interest points within the image pyramid, and extracting feature descriptors for each respective interest point. According to some example embodiments, the feature descriptors are rotation invariant. Further, the example method may also include tracking movement by matching the feature descriptors to feature descriptors of a previous frame and performing recognition of an object within the image frame based on the feature descriptors. Related example methods and example apparatuses are also provided.

Claims

1. A method comprising:

generating an image pyramid of an image frame;

detecting a plurality of interest points within the image pyramid; and

extracting, via a processor, feature descriptors for each respective interest point.

2. The method of claim 1, wherein extracting the feature descriptors includes extracting the feature descriptors that are rotation invariant.

3. The method of claim 1, further comprising:

tracking movement by matching the feature descriptors to feature descriptors of a previous frame; and

performing recognition of an object within the image frame based on the feature descriptors.

4. The method of claim 1, wherein extracting the feature descriptors includes rotating gradients of an interest point by a given angle to yield a radial gradient transform.

5. The method of claim 1, wherein extracting the feature descriptors includes performing gradient binning or spatial binning.

6. The method of claim 1, wherein extracting the feature descriptors includes defining approximate basis vectors and utilizing the approximate basis vectors to generate an approximate radial gradient transform.

7. The method of claim 1, further comprising selecting at least one of the feature descriptors for matching based on a threshold minimum distance between the feature descriptors.

8. An apparatus comprising at least one processor and at least one memory including computer program code, the at least one memory and the computer program code configured to, with the at least one processor, direct the apparatus at least to:

generate an image pyramid of an image frame;

detect a plurality of interest points within the image pyramid; and

extract feature descriptors for each respective interest point.

9. The apparatus of claim 8, wherein the apparatus directed to extract the feature descriptors includes being directed to extract the feature descriptors that are rotation invariant.

10. The apparatus of claim 8, wherein the apparatus is further directed to:

track movement by matching the feature descriptors to feature descriptors of a previous frame; and

perform recognition of an object within the image frame based on the feature descriptors.

11. The apparatus of claim 8, wherein the apparatus directed to extract the feature descriptors includes being directed to rotate gradients of an interest point by a given angle to yield a radial gradient transform.

12. The apparatus of claim 8, wherein the apparatus directed to extract the feature descriptors includes being directed to perform gradient binning or spatial binning.

13. The apparatus of claim 8, wherein the apparatus directed to extract the feature descriptors includes being directed to define approximate basis vectors and utilize the approximate basis vectors to generate an approximate radial gradient transform.

14. The apparatus of claim 8, wherein the apparatus is further directed to select at least one of the feature descriptors for matching based on a threshold minimum distance between the feature descriptors.

15. The apparatus of claim 8, wherein the apparatus comprises a mobile terminal.

16. The apparatus of claim 15, wherein the apparatus includes a display configured to present object tracking based on the feature descriptors.

17. A non-transitory computer-readable storage medium having executable computer-readable program code instructions stored therein, the instructions being configured to cause an apparatus at least to:

generate an image pyramid of an image frame;

detect a plurality of interest points within the image pyramid; and

extract feature descriptors for each respective interest point.

18. The storage medium of claim 17, wherein the instructions configured to cause the apparatus to extract the feature descriptors include being configured to cause the apparatus to extract the feature descriptors that are rotation invariant.

19. The storage medium of claim 17, wherein the instructions are further configured to cause the apparatus to:

track movement by matching the feature descriptors to feature descriptors of a previous frame; and

perform recognition of an object within the image frame based on the feature descriptors.

20. The storage medium of claim 17, wherein the instructions configured to cause the apparatus to extract the feature descriptors include being configured to cause the apparatus to rotate gradients of an interest point by a given angle to yield a radial gradient transform.