SYSTEM AND METHOD ASSOCIATED WITH OBJECT VERIFICATION

• US 20180129868A1
• Filed: 11/08/2016
• Published: 05/10/2018
• Est. Priority Date: 11/08/2016
• Status: Abandoned Application

• Current Assignee: Qualcomm, Inc.
• Original Assignee: Qualcomm, Inc.
• Inventors: Sarkis, Michel Adib, Qi, Yingyong
• Application Number:
  US15/346,343
• Publication Number:
  US 20180129868A1
• Time in Patent Office: Days
• CPC Class Codes: G06F 18/22 Matching criteria, e.g. proximity measures
• CPC Class Codes: G06F 18/25 Fusion techniques
• CPC Class Codes: G06V 10/761 Proximity, similarity or dissimilarity measures
• CPC Class Codes: G06V 10/80 Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level multimodal speaker identification or verification G10L17/10
• CPC Class Codes: G06V 40/168 Feature extraction; Face representation
• CPC Class Codes: G06V 40/172 Classification, e.g. identification

Abstract

An apparatus includes a processor and a memory. The processor is configured to execute instructions stored at the memory to receive first characterization data and second characterization data. The first characterization data includes first values in a first order and corresponding to a first object. The second characterization data includes second values in a second order and corresponding to a second object. The processor is further configured to generate third characterization data and to generate fourth characterization. The third characterization data includes the first values in a third order. The fourth characterization data includes the second values in a fourth order. The processor is also configured to perform a first similarity operation using the first, second, third, and fourth characterization data to generate first result data and to determine whether the first object and the second object match based on the first result data.

Claims

1. An apparatus comprising:

processor; and

a memory coupled to the processor, the memory comprising instructions that, when executed by the processor, cause the processor to;

receive first characterization data and second characterization data, the first characterization data including a plurality of first values in a first order and corresponding to a first object associated with a first image, the second characterization data including a plurality of second values in a second order and corresponding to a second object associated with a second image;

generate third characterization data based on the first characterization data, the third characterization data having the plurality of first values in a third order;

generate fourth characterization data based on the second characterization data, the fourth characterization data having the plurality of second values in a fourth order;

perform a first similarity operation using the first characterization data, the second characterization data, the third characterization data, and the fourth characterization data to generate first result data; and

determine whether the first object and the second object match based on the first result data.

2. The apparatus of claim 1, wherein the first object comprises a first face, wherein the second object comprises a second face, wherein the first characterization data comprises first feature values associated with landmarks of the first face, wherein the second characterization data comprises second feature values associated with landmarks of the second face, and wherein the landmarks of the first face and the landmarks of the second face include a nose, a mouth, a right eye, a left eye, a chin, or a combination thereof.

3. The apparatus of claim 1, wherein the first characterization data corresponds to a first set of overlapping patches of the first image, and wherein the second characterization data correspond to a second set of overlapping patches of the second image.

4. The apparatus of claim 1, wherein the metric data is determined based on a multi-layer quasi-gradient boost process.

5. The apparatus of claim 1, wherein each of the first characterization data, the second characterization data, the third characterization data, and the fourth characterization data comprises a vector, wherein the third order is a reverse order of the first order, and wherein the fourth order is a reverse order of the second order.

6. The apparatus of claim 1, wherein the processor is further configured to:

perform a second similarity operation using the first characterization data, the second characterization data, the third characterization data, and the fourth characterization data to generate second result data; and

combine the first result data and the second result data to generate similarity data.

7. The apparatus of claim 6, wherein the first operation comprises an element-wise absolute difference operation, and wherein the second operation comprises an element-wise multiplication operation.

8. The apparatus of claim 6, wherein the processor is further configured to calculate one or more dot product operations between the similarity data and metric data to generate a similarity score.

9. The apparatus of claim 8, wherein the processor is further configured to compare the similarity score to a threshold and to determine that the first object and the second object are the same object based on the similarity score satisfying the threshold.

10. The apparatus of claim 1, further comprising:

an antenna; and

a receiver coupled to the antenna and to the processor, wherein the memory, the processor, the receiver, and the antenna are integrated into a mobile communication device.

11. A method of performing object verification, the method comprising:

receiving first characterization data and second characterization data, the first characterization data including a plurality of first values in a first order and corresponding to a first object associated with a first image, the second characterization data including a plurality of second values in a second order and corresponding to a second object associated with a second image;

generating third characterization data based on the first characterization data, the third characterization data having the plurality of first values in a third order;

generating fourth characterization data based on the second characterization data, the fourth characterization data having the plurality of second values in a fourth order;

performing a first similarity operation using the first characterization data, the second characterization data, the third characterization data, and the fourth characterization data to generate first result data; and

determining whether the first object and the second object match based on the first result data.

12. The method of claim 11, further comprising determining the metric data using a multi-layer quasi-gradient boost process.

13. The method of claim 11, wherein the third order is a reverse order of the first order, and wherein the fourth order is a reverse order of the second order.

14. The method of claim 11, further comprising:

performing a second similarity operation using the first characterization data, the second characterization data, the third characterization data, and the fourth characterization data to generate second result data; and

combining the first result data and the second result data to generate similarity data.

15. The method of claim 14, wherein the similarity data comprises a similarity vector that includes the first result data and the second result data, and wherein combining the first result data and the second result data comprises concatenating the first result data and the second result data to form the similarity vector.

16. The method of claim 15, wherein the first operation comprises an element-wise absolute difference operation, and wherein the second operation comprises an element-wise multiplication operation.

17. The method of claim 14, further comprising:

calculating one or more dot product operations between the similarity data and metric data to generate a similarity score; and

comparing the similarity score to a threshold, the first object and the second object are determined to be the same object based on the similarity score satisfying the threshold.

18. The method of claim 17, wherein the similarity score satisfies the threshold when the similarity score is greater than or equal to the threshold, and further comprising outputting an indication of whether the first object matches the second object.

19. The method of claim 17, further comprising:

receiving the metric data; and

storing the metric data at a memory.

20. The method of claim 17, further comprising generating the metric data.

21. An apparatus comprising:

means for receiving first characterization data and second characterization data, the first characterization data including a plurality of first values in a first order and corresponding to a first object associated with a first image, the second characterization data including a plurality of second values in a second order and corresponding to a second object associated with a second image;

means for generating third characterization data based on the first characterization data and for generating fourth characterization data based on the second characterization data, the third characterization data having the plurality of first values in a third order, the fourth characterization data having the plurality of second values in a fourth order;

means for performing a first similarity operation using the first characterization data, the second characterization data, the third characterization data, and the fourth characterization data to generate first result data; and

means for determining whether the first object and the second object match based on the first result data.

22. The apparatus of claim 21, wherein the first object comprises a first face, wherein the second object comprises a second face, wherein the first characterization data comprises first feature values associated with landmarks of the first face, wherein the second characterization data comprises second feature values associated with the landmarks of the second face, and wherein the landmarks of the first face and the landmarks of the second face include a nose, a mouth, a right eye, a left eye, a chin, or a combination thereof.

23. The apparatus of claim 21, wherein the first characterization data corresponds to a first set of overlapping patches of the first image, and wherein the second characterization data correspond to a second set of overlapping patches of the second image.

24. The apparatus of claim 21, wherein the means for receiving the first characterization data and the second characterization data, the means for generating the third characterization data and generating the fourth characterization data, the means for performing the first similarity operation, and the means for determining are integrated into a mobile phone, a cellular phone, a computer, a portable computer, a tuner, a radio, a satellite radio, a communication device, a modem, a portable music player, a portable digital video player, a navigation device, a personal digital assistant (PDA), a mobile location data unit, or a combination thereof.

25. A non-transitory computer-readable medium comprising instructions that, when executed by a processor, cause the processor to:

receive first characterization data and second characterization data, the first characterization data including a plurality of first values in a first order and corresponding to a first object associated with a first image, the second characterization data including a plurality of second values in a second order and corresponding to a second object associated with a second image;

generate third characterization data based on the first characterization data, the third characterization data having the plurality of first values in a third order;

generate fourth characterization data based on the second characterization data, the fourth characterization data having the plurality of second values in a fourth order;

perform a first similarity operation using the first characterization data, the second characterization data, the third characterization data, and the fourth characterization data to generate first result data; and

determine whether the first object and the second object match based on the first result data.

26. The non-transitory computer-readable medium of claim 25, wherein the instructions, when executed by the processor, further cause the processor to:

perform a second similarity operation using the first characterization data, the second characterization data, the third characterization data, and the fourth characterization data to generate second result data;

combine the first result data and the second result data to generate similarity data;

calculate one or more dot product operations between the similarity data and metric data to generate a similarity score; and

compare the similarity score to a threshold, the first object and the second object are determined to be the same object based on the similarity score satisfying the threshold.

27. The non-transitory computer-readable medium of claim 26, wherein, to generate the metric data, the instructions further cause the processor to determine the metric data using a multi-layer quasi-gradient boost process.

28. The non-transitory computer-readable medium of claim 26, wherein, to generate the metric data, the instructions further cause the processor to:

generate a similarity matrix based on multiple similarity data sets, wherein each similarity data set of the multiple similarity data sets is generated based on two corresponding training images that include a matching object; and

divide the similarity matrix into multiple portions, each portion of the multiple portions including at least one value from each of the multiple similarity data sets.

29. The non-transitory computer-readable medium of claim 28, wherein the instructions, when executed by the processor, further causes the processor to determine a plurality of portion match scores, the plurality of portion match scores determined by, for each portion, determining a corresponding portion match score based on a match indicator value and the portion.

30. The non-transitory computer-readable medium of claim 29, wherein the instructions, when executed by the processor, further causes the processor to:

determine an average match score of the plurality of portion match scores; and

determine the metric data based on the average match score, the similarity matrix, and the match indicator value.