Cognitive image detection and recognition

• US 10,540,541 B2
• Filed: 05/27/2015
• Issued: 01/21/2020
• Est. Priority Date: 05/27/2014
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Dugan, Casey, Geyer, Werner
• Primary Examiner(s): Perlman, David
• Application Number:
  US14/722,266
• Publication Number:
  US 20150347825A1
• Time in Patent Office: 1,700 Days
• Field of Search: None
• CPC Class Codes: G06F 18/2155 characterised by the incorporation of unlabelled data, e.g. multiple instance learning [MIL], semi-supervised techniques using expectation-maximisation [EM] or naïve labelling
• CPC Class Codes: G06F 18/40 Software arrangements specially adapted for pattern recognition, e.g. user interfaces or toolboxes therefor
• CPC Class Codes: G06V 10/96 Management of image or video recognition tasks
• CPC Class Codes: G06V 40/161 Detection; Localisation; Normalisation
• CPC Class Codes: G06V 40/172 Classification, e.g. identification

Abstract

Embodiments are directed to an image processing station for associating a non-trusted data set with a trusted data set. The station includes a memory and a user interface configured to receive the trusted data set and the non-trusted data set. The station further includes a processor system communicatively coupled to the memory and the user interface. The processor system is configured to generate, based at least in part on processing the trusted data set and the non-trusted data set, a confidence value that represents a probability that at least a portion of the non-trusted data set identifies at least a portion of the trusted data set.

Claims

1. An image processing station for associating a non-trusted data set with a trusted data set, the station comprising;

a user interface configured to receive the non-trusted data set from a user;

the user interface comprising a trusted data source configured to capture the trusted data set;

wherein the non-trusted data set received at the user interface from the user has not been associated with the trusted data set;

wherein the trusted data set captured at the trusted data source of the user interface does not include the non-trusted data set;

a memory; and

a processor system communicatively coupled to the memory and the user interface;

the processor system configured to perform processor operations configured to determine that the trusted data set is associated with the non-trusted data set, the processor operations comprising;

receiving the trusted data set and the non-trusted data set;

analyzing the trusted data set and the non-trusted data set to determine one or more relationships between the trusted data set and the non-trusted data set;

generate, based at least in part on an output from analyzing the trusted data set and the non-trusted data set, a confidence value that represents a probability that at least a portion of the non-trusted data set identifies at least a portion of the trusted data set; and

making a determination that the trusted data set is associated with the non-trusted data set based at least in part on the confidence value;

wherein the processor operations assume that the non-trusted data set is not associated with the trusted data set until the processor operations make the determination that the trusted data set is associated with the non-trusted data set.

2. The station of claim 1, wherein:

the processor operations further comprise, based at least in part on an assessment of the confidence value, storing in a first portion of the memory identified trusted data sets; and

the identified trusted data sets comprise trusted data that has been determined to be identified by non-trusted data.

3. The station of claim 2, wherein the processor operations further comprise generating the confidence value based at least in part on at least one identified data set that was previously generated by the image processing station.

4. The station of claim 3, wherein the processor operations further comprise generating the confidence value based at least in part on at least one identified trusted data set generated by a separate image processing station communicatively coupled to the image processing station through a network.

5. The station of claim 1, wherein:

the processor operations further comprise, based at least in part on the confidence value not exceeding a threshold, storing in a second portion of the memory partially identified trusted data sets; and

the partially identified trusted data sets comprise trusted data that has not been determined to be identified by non-trusted data.

6. The station of claim 5, wherein the processor operations further comprise generating the confidence value based at least in part on at least one partially identified trusted data set that was previously generated by the image processing station.

7. The station of claim 6, wherein the processor operations further comprise generating the confidence value based at least in part on at least one partially identified trusted data set generated by a separate image processing station communicatively coupled to the image processing station through a network.

8. The station of claim 1, wherein the processor system comprises multiple processor circuits.

9. The station of claim 1, wherein:

the portion of the non-trusted data set comprises a name; and

the portion of the trusted data set comprises a face image.

10. The station of claim 1, wherein the processor system comprises:

a comparison and reasoning circuit;

a face recognition circuit; and

a face detection circuit.

11. The station of claim 10, wherein the face detection circuit comprises:

a greedy face detector; and

a conservative face detector.

12. The station of claim 1, wherein the user interface comprises a camera.

13. The station of claim 1, wherein analyzing comprises applying data comparison operations and rule-based reasoning operations to the trusted data set and the non-associated non-trusted data.

14. A computer program product for associating a non-trusted data set with a trusted data set, the computer program product comprising:

a computer readable storage medium having program instructions embodied therewith, wherein the computer readable storage medium is not a transitory signal per se, the program instructions readable by at least one processor circuit of an image processing station to cause the at least one processor circuit to perform a method configured to determine that the trusted data set is associated with the trusted data, the method comprising;

receiving, at a user interface of the image processing station, the non-trusted data set from a user;

capturing, at a trusted data source of the user interface, the trusted data set;

wherein the non-trusted data set received from the user comprises non-associated non-trusted data that has not been associated with the trusted data set;

wherein the trusted data set captured at the trusted data source does not include the non-trusted data set;

receiving the trusted data set and the non-trusted data set;

analyzing, using the at least one processor system of the image processing station communicatively coupled to a memory and the user interface, the trusted data set and the non-trusted data set to determine one or more relationships between the trusted data set and the non-trusted data set; and

generating, using the at least one processor circuit communicatively coupled to the memory and the user interface, based at least in part on an output from analyzing the trusted data set and the non-trusted data set to determine one or more relationships between the trusted data set and the non-trusted data set, a confidence value that represents a probability that at least a portion of the non-trusted data set identifies at least a portion of the trusted data set; and

making a determination that the trusted data set is associated with the non-trusted data set based at least in part on the confidence value;

wherein the method assumes that the non-trusted data set is not associated with the trusted data set until the method makes the determination that the trusted data set is associated with the non-trusted data set.

15. The computer program product of claim 14, wherein the method performed by the at least one processor circuit further comprises:

storing in a first portion of the memory, based at least in part on an assessment of the confidence value, identified trusted data sets; and

storing in a second portion of the memory, based at least in part on the confidence value not exceeding a threshold, partially identified trusted data sets;

wherein, the identified trusted data sets comprise trusted data that has been determined to be identified by non-trusted data; and

wherein, the partially identified trusted data sets comprise trusted data that has not been determined to be identified by non-trusted data.

16. The computer program product of claim 15, wherein generating further comprises generating the confidence value based at least in part on:

at least one identified trusted data set previously generated by the at least one processor circuit; and

at least one identified trusted data set generated by a first separate image processing station communicatively coupled to the image processing station through a network.

17. The computer program product of claim 16, wherein the generating further comprises generating the confidence value based at least in part on:

at least partially identified trusted data set previously generated by the at least one processor circuit; and

at least one partially identified trusted data set generated by a second separate image processing station communicatively coupled to the image processing station through a network.

18. The computer program product of claim 14, wherein analyzing comprises applying data comparison operations and rule-based reasoning operations to the trusted data set and the non-associated non-trusted data.