Methods and systems for generating composite image descriptors

Patent metadata

Field	Value
Publication number	US-12056212-B2
Application number	US-202217897741-A
Country	US
Kind code	B2
Filing date	Aug 29, 2022
Priority date	Feb 4, 2021
Publication date	Aug 6, 2024
Grant date	Aug 6, 2024

Abstract

An illustrative image descriptor generation system determines a subset of image descriptors from a plurality of image descriptors that each correspond to a different feature point included within an image. The subset of image descriptors is determined based on geometric proximity, within the image, of respective feature points of the subset of image descriptors to a feature point of a primary image descriptor. The image descriptor generation system then selects a secondary image descriptor from the subset of image descriptors and combines the primary image descriptor and the secondary image descriptor to form a composite image descriptor. Corresponding methods and systems are also disclosed.

First claim

What is claimed is:

1. A method comprising:
determining a subset of image descriptors from a plurality of image descriptors that each correspond to a different feature point included within an image, the subset of image descriptors determined based on geometric proximity, within the image, of respective feature points of the subset of image descriptors to a feature point of a primary image descriptor, wherein the determining the subset of image descriptors comprises:
generating a geometric map representing the plurality of image descriptors by partitioning the image into a plurality of bins;
assigning each image descriptor of the plurality of image descriptors to a respective bin of the plurality of bins; and
determining, for each particular bin of the plurality of bins that is within a search radius of a first bin to which the primary image descriptor is assigned, whether any of the plurality of image descriptors is assigned to the particular bin;
selecting a secondary image descriptor from the subset of image descriptors; and
combining the primary image descriptor and the secondary image descriptor to form a composite image descriptor.

2. The method of claim 1 , further comprising:
generating a descriptor listing that includes the plurality of image descriptors by:
accessing the image,
analyzing the image to identify feature points included within the image, and
generating, for each of the respective feature points that are identified, a respective image descriptor to be among the plurality of image descriptors included in the descriptor listing; and

identifying, from the descriptor listing, the primary image descriptor.

3. The method of claim 1 , wherein the selecting of the secondary image descriptor includes:
filtering the subset of image descriptors to exclude image descriptors that:
correspond to respective feature points more than a threshold distance away from the feature point of the primary image descriptor within the image, or
are more similar to the primary image descriptor than a similarity threshold; and

designating, as the selected secondary image descriptor, a least similar image descriptor from the filtered subset of image descriptors.

4. The method of claim 1 , further comprising:
selecting one or more additional image descriptors from the subset of image descriptors; and
combining the primary image descriptor with each of the one or more additional image descriptors to form one or more additional composite image descriptors.

5. The method of claim 1 , wherein the selecting of the secondary image descriptor:
includes filtering the subset of image descriptors to exclude image descriptors that correspond to respective feature points more than a threshold distance away from the feature point of the primary image descriptor within the image; and
is based on a geometric proximity between the feature point of the primary image descriptor and a feature point of the secondary image descriptor.

6. The method of claim 1 , wherein the selecting of the secondary image descriptor:
includes filtering the subset of image descriptors to exclude image descriptors that are more similar to the primary image descriptor than a similarity threshold; and
is based on a similarity between the primary image descriptor and the secondary image descriptor.

7. The method of claim 1 , wherein the image depicts a real-world scene and the method further comprises providing the composite image descriptor to an extended reality generation system for use in identifying one or more predetermined target objects depicted within the image.

8. The method of claim 7 , wherein the determining of the subset of image descriptors, the selecting of the secondary image descriptor, the combining of the primary and secondary image descriptors to form the composite image descriptor, and the providing of the composite image descriptor to the extended reality generation system are all performed in real time as the extended reality generation system captures the image and identifies the one or more predetermined target objects depicted within the image.

9. The method of claim 1 , further comprising generating, based on the geometric map, a proximity listing for the primary image descriptor, the proximity listing indicating the subset of image descriptors and generated by including in the proximity listing any of the plurality of image descriptors that is determined to be assigned to any of the particular bins within the search radius of the first bin.

10. The method of claim 1 , wherein the plurality of bins are arranged in horizontal rows and vertical columns, and wherein the number of horizontal rows and vertical columns is proportional to an image resolution of the image.

11. A system comprising:
one or more processors configured to perform a process comprising:
determining a subset of image descriptors from a plurality of image descriptors that each correspond to a different feature point included within an image, the subset of image descriptors determined based on geometric proximity, within the image, of respective feature points of the subset of image descriptors to a feature point of a primary image descriptor, wherein the determining the subset of image descriptors comprises:
generating a geometric map representing the plurality of image descriptors by partitioning the image into a plurality of bins;
assigning each image descriptor of the plurality of image descriptors to a respective bin of the plurality of bins; and
determining, for each particular bin of the plurality of bins that is within a search radius of a first bin to which the primary image descriptor is assigned, whether any of the plurality of image descriptors is assigned to the particular bin;

selecting a secondary image descriptor from the subset of image descriptors; and
combining the primary image descriptor and the secondary image descriptor to form a composite image descriptor.

12. The system of claim 11 , wherein the process further comprises:
generating a descriptor listing that includes the plurality of image descriptors by:
accessing the image,
analyzing the image to identify feature points included within the image, and
generating, for each of the respective feature points that are identified, a respective image descriptor to be among the plurality of image descriptors included in the descriptor listing; and

identifying, from the descriptor listing, the primary image descriptor.

13. The system of claim 11 , wherein the selecting of the secondary image descriptor includes:
filtering the subset of image descriptors to exclude image descriptors that:
correspond to respective feature points more than a threshold distance away from the feature point of the primary image descriptor within the image, or
are more similar to the primary image descriptor than a similarity threshold; and

designating, as the selected secondary image descriptor, a least similar image descriptor from the filtered subset of image descriptors.

14. The system of claim 11 , wherein the process further comprises:
selecting one or more additional image descriptors from the subset of image descriptors; and
combining the primary image descriptor with each of the one or more additional image descriptors to form one or more additional composite image descriptors.

15. The system of claim 11 , wherein the selecting of the secondary image descriptor:
includes filtering the subset of image descriptors to exclude image descriptors that correspond to respective feature points more than a threshold distance away from the feature point of the primary ima

Assignees

• Verizon Patent & Licensing Inc

Inventors

• Virodov Sergey
• Polavarapu Vinaya Kumar

Classifications

• G06V30/18152

  Extracting features based on a plurality of salient regional features, e.g. "bag of words" · CPC title

• G06V10/23

  based on positionally close patterns or neighbourhood relationships · CPC title

• G06V10/46

  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) · CPC title

• G06T2207/20021

  Dividing image into blocks, subimages or windows · CPC title

• G06F18/2113

  by ranking or filtering the set of features, e.g. using a measure of variance or of feature cross-correlation · CPC title