Invariant-based dimensional reduction of object recognition features, systems and methods

Patent metadata

Field	Value
Publication number	US-9792529-B2
Application number	US-201615245986-A
Country	US
Kind code	B2
Filing date	Aug 24, 2016
Priority date	Feb 19, 2014
Publication date	Oct 17, 2017
Grant date	Oct 17, 2017

Abstract

A sensor data processing system and method is described. Contemplated systems and methods derive a first recognition trait of an object from a first data set that represents the object in a first environmental state. A second recognition trait of the object is then derived from a second data set that represents the object in a second environmental state. The sensor data processing systems and methods then identifies a mapping of elements of the first and second recognition traits in a new representation space. The mapping of elements satisfies a variance criterion for corresponding elements, which allows the mapping to be used for object recognition. The sensor data processing systems and methods described herein provide new object recognition techniques that are computationally efficient and can be performed in real-time by the mobile phone technology that is currently available.

First claim

What is claimed is:

1. A sensor data processing system comprising:
a memory; and
one or more image processing computers in communication with the memory, the one or more image processing computers configured to:
obtain a first training data set representative of at least one object related to at least one person under a defined environmental state within a controlled environment, wherein the controlled environment comprises environmental parameters that correspond to environmental attributes;
generate a trait vocabulary as a function of descriptors derived from at least the first training data set;
derive a first recognition trait comprising a first plurality of elements from the first training data set according to a trait extraction algorithm, wherein the first recognition trait is derived using the trait vocabulary;
obtain a second training data set representative of the at least one object related to at least one person under a second environmental state within the controlled environment, wherein the second environmental state is created by adjusting at least one environmental parameter corresponding to an environmental attribute;
derive a second recognition trait comprising a second plurality of elements from the second training data set according to the trait extraction algorithm; and
identify a mapping that maps a plurality of elements of the first recognition trait and the second recognition trait in a new representation space, wherein the mapping of the plurality of elements in the new representation space satisfies trait element variance criteria among corresponding elements in traits across the first training data set and the second training data set.

2. The system of claim 1 , further comprising storing at least one of the trait vocabulary and the mapping in the memory.

3. The system of claim 1 , wherein the second recognition trait has a correspondence to the first recognition trait.

4. The system of claim 1 , wherein the second recognition trait is derived using the trait vocabulary.

5. The system of claim 1 , wherein the trait extraction algorithm comprises an image processing algorithm.

6. The system of claim 5 , wherein the image processing algorithm comprises at least one of a SIFT, FAST, FREAK, BRIEF, ORB, BRISK, GLOH, SURF, vSLAM, SLAM, BURST, and DAISY image processing algorithm.

7. The system of claim 1 , wherein the trait extraction algorithm comprises one or more classification algorithms and object recognition algorithms.

8. The system of claim 1 , wherein the trait extraction algorithm comprises at least one edge-based recognition technique.

9. The system of claim 1 wherein the first recognition trait comprises a descriptor.

10. The system of claim 9 , wherein the descriptor comprises an image descriptor.

11. The system of claim 10 , wherein the first plurality of elements comprises dimensions of the descriptor.

12. The system of claim 1 , wherein the first recognition trait comprises a cluster of descriptors.

13. The system of claim 12 , wherein the cluster of descriptors comprises a constellation of descriptors.

14. The system of claim 13 , wherein the constellation of descriptors is within a descriptor space.

15. The system of claim 13 , wherein the constellation of descriptors is with respect to three-dimensional space.

16. The system of claim 12 , wherein the first plurality of elements comprises at least one of a descriptor, a descriptor location, and a cluster of descriptors.

17. The system of claim 1 , wherein the trait vocabulary comprises a corpus of vocabulary atoms representing traits.

18. The system of claim 17 , wherein at least one of the first recognition trait and the second recognition trait are associated with at least one of the vocabulary atoms.

19. The system of claim 17 , wherein one or more elements of the first recognition trait or second recognition trait comprise a distribution or histogram of vocabulary atoms.

20. The system of claim 17 , wherein the corpus of vocabulary atoms representing traits comprises at least one cluster shape trait.

21. The system of claim 1 , further comprising one or more object categorization computers configured to classify the at least one object related to at least one person as a type of object related to medical information based on at least one of the first recognition trait and the second recognition trait.

22. The system of claim 21 , wherein the one or more object categorization computers are further configured to store the at least one of the first recognition trait and the second recognition trait with the type of object related to medical information in a known object database.

23. The system of claim 21 , wherein the medical information comprises biometric data.

24. The system of claim 1 , wherein at least one of the first training data set and the second training data set is related to medical imaging.

25. The system of claim 1 , wherein the mapping represents an invariant property as a recognition property with respect to the at least one environmental parameter adjusted as exhibited by at least some of the first plurality of elements.

26. The system of claim 1 , wherein the mapping represents a variant property as a recognition property with respect to the at least one environmental parameter adjusted as exhibited by at least some of the first plurality of elements.

27. The system of claim 1 , wherein the mapping comprises a dimensionality reduction of the plurality of elements.

28. The system of claim 1 , wherein the mapping comprises a non-linear mapping from the first plurality of elements and the second plurality of elements to a third plurality of elements in a new multi-dimensional space.

29. The system of claim 1 , wherein the mapping comprises a look-up table.

30. The system of claim 1 , wherein the mapping comprises a linear mapping from the first plurality of elements and the second plurality of elements to a third plurality of elements in a new invariant space.

31. The system of claim 1 , wherein the mapping comprises an inferred state of a real-world environment based on the second recognition trait.

32. The system of claim 1 , wherein adjusting at least one environmental parameter comprises adjusting at least one of a wireless signal, a lighting property, a camera property, a sensor property, an error rate, a gravity field, a shape, a distance, a field-of-view, a scale, a duration or length of time, a sampling or analysis frequency, a distortion of time via slowing down or speeding up sensor data playback, and an orientation.

33. The system of claim 1 , wherein the defined environmental state comprises an expected environmental state external to the controlled environment.

34. The system of claim 1 , wherein the trait element variance criteria identify a low variance among the corresponding elements in the traits across the first training set and the second training set, the low variance operating as a function of a low variance threshold.

35. The system of claim 1 , wherein the trait element variance criteria identify a high variance among the corresponding elements in the traits across the first training set and the second training set, the high variance operating as a function of a high variance threshold

Assignees

• Nant Holdings Ip Llc

Inventors

• Wnuk Kamil
• Sudol Jeremi
• Song Bing
• Siddiqui Matheen
• Mckinnon David

Classifications

• G06V10/771

  Feature selection, e.g. selecting representative features from a multi-dimensional feature space · 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

• G06K9/623

  Physics · mapped topic

• G06K9/4633

  Physics · mapped topic

• G06K9/6232 Primary

  Physics · mapped topic