Scalable and distributed biometric processing

• US 9,779,316 B2
• Filed: 11/05/2015
• Issued: 10/03/2017
• Est. Priority Date: 09/30/2015
• Status: Active Grant

• Current Assignee: Sensormatic Electronics LLC (Johnson Controls International plc)
• Original Assignee: Tyco Fire & Security GmbH (Johnson Controls International plc)
• Inventors: Rao, Manjuprakash R., Borah, Surajit, Ramanna, Sreenath K., Chinta, Rambabu
• Primary Examiner(s): Dang, Duy M
• Application Number:
  US14/933,712
• Publication Number:
  US 20170091570A1
• Time in Patent Office: 698 Days
• CPC Class Codes: G06F 18/217 Validation; Performance evaluation; Active pattern learning techniques
• CPC Class Codes: G06V 10/95 structured as a network, e.g. client-server architectures
• CPC Class Codes: G06V 40/162 using pixel segmentation or colour matching
• CPC Class Codes: G06V 40/172 Classification, e.g. identification

Abstract

Processing system and method for biometrics identification uses a plurality of biometrics capture devices and one or more network servers to provide cloud based biometric processing services. One or more gateway computers regulate data traffic as between the one or more servers and the biometric capture devices. The components of the system work cooperatively to reduce system bandwidth requirements and reduce a processing load upon various system components. The components are responsive to processing requests which dynamically selectively specify constituent steps in a biometric recognition process which are to be performed by each of the components.

Claims

1. A distributed processing system for biometrics identification, comprising:

a plurality of biometrics capture devices disposed at a plurality of biometric capture locations and connected to at least a first computer data network;

at least one server remote from the biometric capture devices which processes biometric feed data from the plurality of biometric capture devices, the at least one server connected to a second computer data network;

at least one gateway connected to the first and second computer data networks, the at least one gateway regulating data traffic as between the at least one server and the plurality of biometric capture devices;

the plurality of biometric capture devices, the at least one gateway, and the at least one server cooperatively arranged to perform processing operations and communications operations to execute a biometric recognition process in which an individual person is recognized based on the biometric data captured by the biometric capture devices;

wherein components of the distributed processing system including the at least one gateway and the plurality of biometric capture devices are each responsive to a plurality of processing requests, said processing requests dynamically selectively specifying constituent steps in a biometric recognition process instance which are to be performed by each of the components for carrying out the biometric recognition process; and

wherein the constituent steps performed by the plurality of biometric capture devices involve processing captured biometric data to (a) detect biometric features associated with at least one person and (b) reduce an amount of the captured biometric data that is to be communicated over at least the first computer data network based on the detected biometric features.

2. The distributed processing system according to claim 1, wherein the processing requests directed to the plurality of biometric capture devices are coordinated by a processing optimizer module disposed in the at least one gateway.

3. The distributed processing system according to claim 2, wherein the processing requests directed to each of said components are determined by the processing optimizer module in accordance with the capabilities of the components and the processing requirements for the core functions of the components.

4. The distributed processing system according to claim 3, wherein the execution of the constituent steps of each biometric recognition process instance are adaptively varied by the processing optimizer module among the components available in a network neighborhood based on information concerning the processing load of the components.

5. The distributed processing system according to claim 4, wherein the processing optimizer module applies a set of assigned rules to determine an optimal selection of the component in the network neighborhood for execution of the constituent steps.

6. The distributed processing system according to claim 5, wherein the set of assigned rules is dynamically controlled by the server and/or the gateway based on at least one machine learning algorithm.

7. The distributed processing system according to claim 4, wherein the plurality of biometric capture devices periodically update the processing optimizer module with information concerning processing resource availability or load.

8. The distributed processing system according to claim 1, wherein performance of the biometric recognition process is optimized at the server by adaptive adjustment of threshold values applied when analyzing biometric data that has been captured.

9. The distributed processing system according to claim 1, wherein the plurality of biometric capture devices are each configured to selectively process captured biometric data to limit data communicated to the gateway when biometric recognition of an entity associated with the biometric data has already been established to a predetermined level of confidence.

10. The distributed processing system according to claim 9, wherein the biometric capture devices receive feedback from at least one of the gateway and the server to determine when recognition of an entity associated with the biometric data has been established to the predetermined level of confidence.

11. The distributed processing system according to claim 1, wherein at least one of the gateway and the plurality of biometric capture devices perform at least one processing function for determining biometric data to be communicated, the at least one processing function selected from the group consisting of data quality assessment, data redundancy assessment, data defect assessment, and data relevance assessment.

12. The distributed processing system according to claim 11, wherein the at least one processing function involves an assessment relative to at least one threshold value or combination of multiple threshold values which is/are dynamically determined in accordance with a machine learning algorithm.

13. The method according to claim 11, wherein at least one of the gateway and the plurality of biometric capture devices receive feedback from the at least one server to determine whether the biometric data should be communicated.

14. The distributed processing system according to claim 1, wherein at least one of the gateway and the plurality of biometric capture devices performs at least one processing step to dynamically determine a priority of biometric data to be communicated.

15. The distributed processing system according to claim 14, wherein the priority is based on at least one factor selected from the group consisting of data quality, data redundancy, data defect and data relevance.

16. The distributed processing system according to claim 15, wherein the at least one factor is evaluated in accordance with a dynamic threshold determined in accordance with a machine learning algorithm.

17. The distributed processing system according to claim 1, further comprising:

a database containing biometric data, the database indexed based on a plurality of gross level categories associated with the biometric data;

wherein at least one of the gateway and the plurality of biometric capture devices extracts metadata from the biometric data which has been collected, the metadata containing information indicating one or more of the plurality of gross level categories applicable to the biometric data that has been collected; and

wherein the server is responsive to the gross level categories specified in the metadata to expedite a comparison of the biometric data which has been collected to the biometric data stored in the database.

18. The distributed processing system according to claim 1, wherein one or more of the components operates according to a set of assigned rules which are adaptively determined by the at least one server based on at least one machine learning algorithm.

19. The distributed processing algorithm according to claim 18, wherein the machine learning algorithm uses multi-modal biometrics or other non-biometric factors to create an image database that is used for fine tuning the set of assigned rules.

20. A distributed processing system for biometrics identification, comprising:

a plurality of biometrics capture devices disposed at a plurality of biometric capture locations and connected to at least a first computer data network;

at least one server remote from the biometric capture devices which processes biometric feed data from the plurality of biometric capture devices, the at least one server connected to a second computer data network;

at least one gateway connected to the first and second computer data networks, the at least one gateway regulating data traffic as between the at least one server and the plurality of biometric capture devices;

the plurality of biometric capture devices, the at least one gateway, and the at least one server cooperatively arranged to perform processing operations and communications operations to execute a biometric recognition process in which an individual person is recognized based on the biometric data captured by the biometric capture devices; and

wherein components of the distributed processing system including the at least one gateway and the plurality of biometric capture devices are each responsive to a plurality of processing requests, said processing requests dynamically selectively specifying constituent steps in a biometric recognition process instance which are to be performed by each of the components for carrying out the biometric recognition process;

wherein the processing requests directed to the plurality of biometric capture devices are coordinated by a processing optimizer module disposed in the at least one gateway;

wherein the processing requests directed to each of said components are determined by the processing optimizer module in accordance with the capabilities of the components and the processing requirements for the core functions of the components;

wherein the execution of the constituent steps of each biometric recognition process instance are adaptively varied by the processing optimizer module among the components available in a network neighborhood based on information concerning the processing load of the components;

wherein the processing optimizer module applies a set of assigned rules to determine an optimal selection of the component in the network neighborhood for execution of the constituent steps; and

wherein the optimal selection is determined based on a log which is dynamically maintained by the processing optimizer module and specifying processing requests assigned to each of the available components in the network neighborhood.

21. A method for implementing a scalable distributed processing system for biometrics identification, comprising:

disposing a plurality of biometric capture devices at a plurality of biometric capture locations and connecting them to at least a first computer data network;

disposing at least one server remote from the biometric capture devices to process biometric feed data received from the plurality of biometric capture devices, and connecting the at least one server to a second computer data network;

connecting at least one gateway to the first and second computer data networks, and using the at least one gateway to regulate data traffic as between the at least one server and the plurality of biometric capture devices;

performing cooperative processing and communications operations involving the plurality of biometric capture devices, the at least one gateway, and the at least one server to execute a biometric recognition process in which an individual person is recognized based on the biometric data captured by the biometric capture devices; and

communicating a plurality of processing requests from a processing optimizer module to components of the distributed processing system including the at least one gateway and the plurality of biometric capture devices, and dynamically selectively specifying which constituent steps of an instance of the biometric recognition process associated with each of the processing requests are to be performed by each of the components for carrying out the biometric recognition process;

wherein the constituent steps performed by the plurality of biometric capture devices involve processing captured biometric data to (a) detect biometric features associated with at least one person and (b) reduce an amount of the captured biometric data that is to be communicated over at least the first computer data network based on the detected biometric features.