METHODS AND SYSTEMS FOR DATA ANALYSIS AND FEATURE RECOGNITION

• US 20120141021A1
• Filed: 02/28/2011
• Published: 06/07/2012
• Est. Priority Date: 03/23/2006
• Status: Active Grant

• Current Assignee: IntelliScience Corp.
• Original Assignee: IntelliScience Corp.
• Inventors: Brinson, Robert M. Jr., Middleton, Nicholas L., Donaldson, Bryan G.
• Granted Patent: US 8,625,885 B2
• Time in Patent Office: Days
• US Class Current: 382/160
• CPC Class Codes: G06F 18/24765 Rule-based classification
• CPC Class Codes: G06F 18/41 Interactive pattern learning with a human teacher
• CPC Class Codes: G06V 10/765 using rules for classification or partitioning the feature space
• CPC Class Codes: G06V 10/7788 the supervisor being a human, e.g. interactive learning with a human teacher
• CPC Class Codes: G06V 10/95 structured as a network, e.g. client-server architectures
• CPC Class Codes: G06V 20/13 Satellite images

Abstract

Systems and methods for automated pattern recognition and object detection. The method can be rapidly developed and improved using a minimal number of algorithms for the data content to fully discriminate details in the data, while reducing the need for human analysis. The system includes a data analysis system that recognizes patterns and detects objects in data without requiring adaptation of the system to a particular application, environment, or data content. The system evaluates the data in its native form independent of the form of presentation or the form of the post-processed data.

Claims

1. A system for data analysis and feature recognition comprising:

a computer processor;

a display in communication with the computer processor;

a data store in communication with the computer processor, the data store containing a series of algorithms; and

a memory in communication with the computer processor and containing stored programming instructions operable by the computer processor, the stored programming instruction, when operated by the computer processor, causing the computer processor to;

TRAIN FOR THE PRESENCE OF THE FEATURE WITHIN A FIRST DIGITAL DATA SET BASED ON THE SERIES OF ALGORITHMS, COMPRISING;

(i) calculating a first value for a first target data element within the first digital data set using a first algorithm from the series of algorithms;

(ii) calculating a second value for the first target data element within the first digital data set using a second algorithm from the series of algorithms;

(iii) repeating the steps of calculating the first value and calculating the second value for a plurality of additional target data elements within the first digital data set using the first algorithm and the second algorithm from the series of algorithms;

(iv) defining a plurality of synaptic paths, each one of the plurality of synaptic paths being defined by the calculated first values and calculated second values for each of the target data elements; and

(v) associating the feature with one or more of the synaptic paths; and

IDENTIFY THE FEATURE IN A SECOND DIGITAL DATA SET BASED ON THE SERIES OF ALGORITHMS, COMPRISING;

(vi) calculating a first value for a first target data element within the second digital data set using the first algorithm from the series of algorithms;

(vii) calculating a second value for the first target data element within the second digital data set using the second algorithm from the series of algorithms;

(viii) repeating the steps of calculating the first value and calculating the second value for a plurality of additional target data elements within the second digital data set using the first algorithm and the second algorithm from the series of algorithms; and

(ix) determining whether the feature is present in the second digital data set by comparing the calculated first values and second values from the second digital data set with the defined plurality of synaptic paths associated with the feature.

2. They system of claim 1, wherein:

the stored programming instructions causing the computer processor to train for the presence of the feature further cause the computer processor to;

(i) calculate a third value for the first target data element within the first digital data set using a third algorithm from the series of algorithms;

(ii) repeat the step of calculating the third value for the plurality of additional target data elements within the first digital data set; and

(iii) define the plurality of synaptic paths by the calculated first values, calculated second values, and calculated third values for each of the target data elements;

and further wherein the stored programming instructions causing the computer processor to identify the feature in the second digital data set cause the processor to;

(i) calculate a third value for the first target data element within the second digital data set using the third algorithm from the series of algorithms;

(ii) repeat the step of calculating the third value for the plurality of additional target data elements within the first digital data set; and

(iii) determine whether the feature is present in the second digital data set by comparing the calculated first values, calculated second values, and calculated third values from the second digital data set with the defined plurality of synaptic paths associated with the feature.

3. They system of claim 1, wherein the first digital data set is a multispectral data set of a first type and the second digital data set is a multispectral data set of the first type.

4. They system of claim 3, wherein the first target data element, the second target data element, and the plurality of additional target data elements from the first digital data set are each contained within a common band within the first digital data set.

5. They system of claim 2, wherein the stored programming instructions further cause the processor to associate the feature with one or more of the synaptic paths based upon a cluster value, the cluster value representing a number of times the feature must occur within a defined cluster area for the feature to be associated with the target data element.

6. They system of claim 5, wherein the defined cluster extends across multiple bands within the multispectral data set.

7. They system of claim 1, wherein the stored programming instructions further cause the processor to create a known feature data store, the known feature data store containing unique synaptic paths associated with the feature.

8. They system of claim 7, wherein the known feature data store further comprises a count value for each of a plurality of unique synaptic paths, the count value indicating a number of instances in which the feature has been found to be associated with each one of the unique synaptic paths.

9. They system of claim 1, wherein the synaptic paths comprise a plurality of nodes, each of the nodes representing a calculated value from a different one of the series of algorithms.

10. They system of claim 1, wherein the stored programming instructions further cause the processor to perform an action comprising one of sounding an alarm or displaying an indicator when the processor determines that the feature is present in the second digital data set.

11. They system of claim 1, wherein the first digital data set is a multispectral data set comprising a plurality of subsets of sensor data, the plurality of subsets of sensor data comprising contiguous bands within a range of a frequency spectrum.

12. They system of claim 11, wherein the range of the frequency spectrum comprises the visible/infrared portion of the frequency spectrum, and further wherein the plurality of subsets of sensor data lie within a plurality of different bands within the visible/infrared portion of the frequency spectrum.

13. They system of claim 11, wherein each of the synaptic paths comprises a plurality of nodes, the plurality of nodes representing calculated values corresponding to target data elements from a plurality of subsets of sensor data.

14. A system for data analysis and feature recognition comprising:

a computer processor;

a display in communication with the computer processor;

a data store in communication with the computer processor, the data store containing a series of algorithms;

a first digital data set in communication with the processor, the first digital data set comprising a plurality of subsets, each of the plurality of subsets being drawn from a different band of a frequency spectrum;

a second digital data set in communication with the processor, the second digital data set comprising a plurality of subsets, each of the plurality of subsets being drawn from a different band of a frequency spectrum; and

a memory in communication with the computer processor and containing stored programming instructions operable by the computer processor, the stored programming instruction, when operated by the computer processor, causing the computer processor to;

TRAIN FOR THE PRESENCE OF THE FEATURE WITHIN THE FIRST DIGITAL DATA SET BASED ON THE SERIES OF ALGORITHMS, COMPRISING;

(x) calculating a first value for a first target data element within a first subset from the plurality of subsets within the first digital data set using one of the algorithms from the series of algorithms;

(xi) calculating a second value for a second target data element within a second subset from the plurality of subsets within the first digital data set using one of the algorithms from the series of algorithms;

(xii) repeating the steps of calculating the first value and calculating the second value for a plurality of additional target data elements within first subset and the second subset from the first digital data set;

(xiii) defining a plurality of synaptic paths, each one of the plurality of synaptic paths being defined by the calculated first values and calculated second values for each of the target data elements; and

(xiv) associating the feature with one or more of the synaptic paths; and

IDENTIFY THE FEATURE IN THE SECOND DIGITAL DATA SET BASED ON THE SERIES OF ALGORITHMS, COMPRISING;

(xv) calculating a first value for a first target data element within a first subset from the plurality of subsets within the second digital data set using one of the algorithms from the series of algorithms;

(xvi) calculating a second value for a second target data element within a second subset from the plurality of subsets within the second digital data set using one of the algorithms from the series of algorithms;

(xvii) repeating the steps of calculating the first value and calculating the second value for a plurality of additional target data elements within the second digital data set using the first algorithm and the second algorithm from the series of algorithms; and

(xviii) determining whether the feature is present in the second digital data set by comparing the calculated first values and second values from the second digital data set with the defined plurality of synaptic paths associated with the feature.

15. The system of claim 14 wherein each of the synaptic paths comprises a plurality of nodes, the plurality of nodes representing calculated values corresponding to target data elements from a plurality of subsets of sensor data.

16. They system of claim 14, wherein:

the stored programming instructions causing the computer processor to train for the presence of the feature further cause the computer processor to;

(i) calculate a third value for a third target data element within a third subset from the plurality of subsets within the first digital data set using one of the algorithms from the series of algorithms;

(ii) repeat the step of calculating the third value for the plurality of additional target data elements within the first digital data set; and

(iii) define the plurality of synaptic paths by the calculated first values, calculated second values, and calculated third values for each of the target data elements;

and further wherein the stored programming instructions causing the computer processor to identify the feature in the second digital data set cause the processor to;

(i) calculate a third value for a third target data element within a third subset from the plurality of subsets within the second digital data set using one of the algorithms from the series of algorithms;

(ii) repeat the step of calculating the third value for the plurality of additional target data elements within the first digital data set; and

(iii) determine whether the feature is present in the second digital data set by comparing the calculated first values, calculated second values, and calculated third values from the second digital data set with the defined plurality of synaptic paths associated with the feature.

17. The system of claim 16 wherein the first value, the second value, and the third value are each calculated by using the same algorithm from the series of algorithms.

18. The system of claim 16, wherein the first value, the second value, and the third value are each calculated by using a different algorithm from the series of algorithms.