Media fingerprinting and identification system
The overall architecture and details of a scalable video fingerprinting and identification system that is robust with respect to many classes of video distortions is described. In this system, a fingerprint for a piece of multimedia content is composed of a number of compact signatures, along with traversal hash signatures and associated metadata. Numerical descriptors are generated for features found in a multimedia clip, signatures are generated from these descriptors, and a reference signature database is constructed from these signatures. Query signatures are also generated for a query multimedia clip. These query signatures are searched against the reference database using a fast similarity search procedure, to produce a candidate list of matching signatures. This candidate list is further analyzed to find the most likely reference matches. Signature correlation is performed between the likely reference matches and the query clip to improve detection accuracy.
1. A method comprising:
identifying an initial position of an interest region within a frame of a video sequence;
identifying an initial scale of the interest region;
refining the initial position and the initial scale using an array of anisotropic filters so as to obtain a refined position of the interest region and a refined scale of the interest region;
establishing a pixel region within the frame that is centered at the refined position and has a size in pixels that is proportional to the refined scale;
determining a multi-dimensional signature based on pixels in the pixel region; and
storing the multi-dimensional signature in a database in association with the video sequence.
2. The method of claim 1 , wherein the initial scale defines a spatial extent of the array of anisotropic filters.
3. The method of claim 1 , wherein:
the pixel region is a rectangular box, and
horizontal vertices and vertical vertices of the rectangular box are based on the refined scale.
4. The method of claim 1 , wherein identifying the initial position and the initial scale comprises identifying the initial position and the initial scale using a bi-level filter.
5. The method of claim 4 , wherein the bi-level filter comprises a Laplacian of Gaussian second order partial derivative bi-level filter.
6. The method of claim 1 , wherein the array of anisotropic filters is an array of elliptic-shaped anisotropic filters.
7. The method of claim 1 , wherein determining the multi-dimensional signature comprises:
determining sub-region pixel values based on values of pixels within respective sub-regions of the pixel region; and
determining the multi-dimensional signature using the sub-region pixel values.
8. A computing system comprising:
a processing unit; and
a non-transitory computer-readable medium having stored therein instructions that are executable to cause the processing unit to perform functions comprising:
identifying an initial position of an interest region within a frame of a video sequence,
identifying an initial scale of the interest region,
refining the initial position and the initial scale using an array of anisotropic filters so as to obtain a refined position of the interest region and a refined scale of the interest region,
establishing a pixel region within the frame that is centered at the refined position and has a size in pixels that is proportional to the refined scale,
determining a multi-dimensional signature based on pixels in the pixel region, and
storing the multi-dimensional signature in a database in association with the video sequence.
9. The computing system of claim 8 , wherein the initial scale defines a spatial extent of the array of anisotropic filters.
10. The computing system of claim 8 , wherein:
the pixel region is a rectangular box, and
horizontal vertices and vertical vertices of the rectangular box are based on the refined scale.
11. The computing system of claim 8 , wherein identifying the initial position and the initial scale comprises identifying the initial position and the initial scale using a bi-level filter.
12. The computing system of claim 11 , wherein the bi-level filter comprises a Laplacian of Gaussian second order partial derivative bi-level filter.
13. The computing system of claim 8 , wherein the array of anisotropic filters is an array of elliptic-shaped anisotropic filters.
14. The computing system of claim 8 , wherein determining the multi-dimensional signature comprises:
determining sub-region pixel values based on values of pixels within respective sub-regions of the pixel region; and
determining the multi-dimensional signature using the sub-region pixel values.
15. A non-transitory computer-readable medium having stored therein instructions that are executable to cause a processing unit to perform functions comprising:
identifying an initial position of an interest region within a frame of a video sequence;
identifying an initial scale of the interest region;
refining the initial position and the initial scale using an array of anisotropic filters so as to obtain a refined position of the interest region and a refined scale of the interest region;
establishing a pixel region within the frame that is centered at the refined position and has a size in pixels that is proportional to the refined scale;
determining a multi-dimensional signature based on pixels in the pixel region; and
storing the multi-dimensional signature in a database in association with the video sequence.
16. The non-transitory computer-readable medium of claim 15 , wherein the initial scale defines a spatial extent of the array of anisotropic filters.
17. The non-transitory computer-readable medium of claim 15 , wherein:
the pixel region is a rectangular box, and
horizontal vertices and vertical vertices of the rectangular box are based on the refined scale.
18. The non-transitory computer-readable medium of claim 15 , wherein identifying the initial position and the initial scale comprises identifying the initial position and the initial scale using a bi-level filter.
19. The non-transitory computer-readable medium of claim 15 , wherein the array of anisotropic filters is an array of elliptic-shaped anisotropic filters.
20. The non-transitory computer-readable medium of claim 15 , wherein determining the multi-dimensional signature comprises:
determining sub-region pixel values based on values of pixels within respective sub-regions of the pixel region; and
determining the multi-dimensional signature using the sub-region pixel values.