Learning Contour Identification System Using Portable Contour Metrics Derived From Contour Mappings

• US 20160210534A1
• Filed: 11/16/2015
• Published: 07/21/2016
• Est. Priority Date: 11/18/2014
• Status: Active Grant

• Current Assignee: Harry Friedbert Padubrin
• Original Assignee: Harry Friedbert Padubrin
• Inventors: Padubrin, Harry Friedbert
• Granted Patent: US 9,582,739 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 17/40 Data acquisition and logging for input to computer G06F3/00
• CPC Class Codes: G06F 18/213 Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
• CPC Class Codes: G06F 18/214 Generating training patterns; Bootstrap methods, e.g. bagging or boosting
• CPC Class Codes: G06F 18/24323 Tree-organised classifiers
• CPC Class Codes: G06N 7/01 Probabilistic graphical models, e.g. probabilistic networks
• CPC Class Codes: G06T 2207/20081 Training; Learning
• CPC Class Codes: G06T 7/12 Edge-based segmentation
• CPC Class Codes: G06V 10/44 Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
• CPC Class Codes: G06V 10/46 Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features colour feature extraction G06V10/56
• CPC Class Codes: G06V 10/752 Contour matching
• CPC Class Codes: G06V 10/7715 Feature extraction, e.g. by transforming the feature space, e.g. multi-dimensional scaling [MDS]; Mappings, e.g. subspace methods
• CPC Class Codes: G06V 30/19127 Extracting features by transforming the feature space, e.g. multidimensional scaling; Mappings, e.g. subspace methods

Abstract

A system and method that transforms data formats into contour metrics and further transforms each contour of that mapping into contours pattern metric sets so that each metric created has a representation of one level of contour presentation, at each iteration of the learning contour identification system defined herein. This transformation of data instance to contour metrics permits a user to take relevant data of a data set, as determined by a learning contour identification system, to machines of other types and function, for the purpose of further analysis of the patterns found and labeled by said system. The invention performs with data format representations, not limited to, signals, images, or waveform embodiments so as to identify, track, or detect patterns of, amplitudes, frequencies, phases, and density functions, within the data case and then by way of using combinations of statistical, feedback adaptive, classification, training algorithm metrics stored in hardware, identifies patterns in past data cases that repeat in future, or present data cases, so that high-percentage labeling and identification is a achieved.

Claims

1. ) A computer implemented method for identifying contour groupings, within contour maps, and within at least one learning contour identification system, comprising:

prepare at least one learning contour identification system for processing data types that are internal, and retrieving data types that are both internal and external, with file type format being external containers of data format described by data format in information technology, and where reading data types of whether data recalled was from internal or external format of the data type is dependent upon what stage the learning contour system resides in method execution,

provide training cases of data instances of format numerical data type for at least one learning contour identification system iteratively reading and processing same, or converting said case to a system readable plurality of formatted data types for same system purpose,

transform at least one of the training cases into at least one contour map, of at least one contour, with each contour of the mapping further transformed into having at least one training contour pattern metric set, each defined entirely between two memory addresses when stored, with each contour a contour pattern metric set containing a possibility of at least one;

plurality label sets, plurality coordinate point sets, plurality statistical outcome point sets, plurality calculated outcome point-sets, plurality metric instruction code-sets, and plurality of grouping contours and mappings and their sub-pattern metric sets of same,

store and label each metric of each contour into individual memory addressed locations, wherein managing appending to and removal from the memory being as determined necessary by at least one learning contour identification system'"'"'s pattern identification process,

retrieve from memory, iteratively, a portion of the total finite set of stored training contour pattern metric sets, each training contour pattern metric set retrieved for the purpose of grouping contour pattern metric sets for determining a black boxed or rule-based machine instruction code set, for the classifier of at least one learning contour identification system, that when the instruction code set is tested against the remaining set of labeled and known training contour pattern metric sets, a desired level of performance presented by a confusion matrix is achieved,

store instruction code set and label as a black boxed or rule-based learned instruction set sequence, and store confusion matrix values,

provide test cases of data instances of format numerical data type for at least one learning contour identification system iteratively reading and processing same, or converting said case to a system readable plurality of formatted data types for same system purpose,

transform at least one of the test cases into at least one contour map, of at least one contour, with each contour of the mapping further transformed into having at least one test contour pattern metric set, each defined entirely between two memory addresses when stored, with each contour a contour pattern metric set containing a possibility of at least one;

plurality label sets, plurality coordinate point sets, plurality statistical outcome point sets, plurality calculated outcome point sets, plurality metric instruction code sets, and plurality of grouping contours and mappings and their sub-pattern metric sets of same,

store and label each metric of each contour into individual memory addressed locations, wherein managing appending to and removal from the memory being as determined necessary by at least one learning contour identification system'"'"'s pattern identification process,

retrieve form memory the black boxed or rule-based labeled instruction code set, determined from the learning contour identification system, and retrieve from memory in an iterative process, test contour pattern metrics, to finalize the identification of the unknown test labeled contour pattern metric set combinations optimized in training and captured in the instruction set used to identify contour pattern of interest,

label at least one matched contour pattern metric set as an data item group of interest and compare performance to confusion matrix performance and repeat training and testing with increases or decreases in the number of contours in either test or training transformations, or both, and stop iterations of increases in contours when maximum percentage of success is achieved based on training confusion matrix performance readings,

output to display interfaces the identification of the test contour pattern of the classifier, and output the success reading for that classification from the confusion matrix along with other information pertinent to understanding output by user.

2. ) The method of claim 1, wherein the data instances of the case comprises:

data provided in at least one industry known data type formats readable by at least one learning contour identification system, and

label changes by at least one learning contour identification system.

Instance values converted to combinations of data type formats readable by at least one learning contour identification system.

3. ) The method of claim 2, wherein the label comprises means for contour metric set identification to at least one learning contour identification system whereby label of unknown identifier is as determined by one or more learning contour training systems to be blank, null, or of same meaning as an unknown label.

4. ) The method of claim 2, wherein label comprises:

means for assigning analyzable storage information identifying the contour of interest and its metrics during on time processing within at least one learning contour identification system,

means to make label identities changeable by at least one learning contour identification system to identify a new contour of pattern interest by its same contour metric, accommodating changes in memory.

5. ) The method of claim 2, wherein readable comprises means for:

binary formatted data conversion to data type formats readable by the leaning contour identification system, and

conversion of formatted machine code to readable code determined readable by at least one learning contour identification system to execute said instructions through said systems interfaces.

6. ) The method of claim 2, wherein data type formatting comprises means for at least one of the following:

case conversion of its data instances of non-numerical format to one of numerical format readable by at least one learning contour identification system,

case conversion of its data instances of compressed formats to non-compressed data type formats readable by at least one learning contour identification system,

case conversion of its non-compressed data instances formats to compressed data type formats readable by at least one learning contour identification system,

case conversion of its non-compressed data instances formats to compressed data type formats readable by at least one learning contour identification system,

case conversion from its analog data type formats to digital data type formats readable by at least one learning contour identification system,

case conversion from its digital data type formats to analog data type formats readable by at least one learning contour identification system,

case conversion from its analog data type formats comprising, data types characteristic to physical recording medium storage and storage formats of recording receiving and transmitting device data types,

case conversion from its digital data type formats comprising a data type characteristic of a physical recording medium storage and storage formats of recording receiving and transmitting device data types,

case conversion of its electric generated signal data type formats to data type formats readable by at least one learning contour identification system,

case conversion of its real-time communication channel data type formats into a data type format readable by at least one learning contour identification system,

means for system initialization of data type primitives that are start and end processing requirements of at least one learning contour identification system.

7. ) The method of claim 6, wherein an analog data type comprises a readable transmission and reception of electrical signal data types translated into electric data types of pulses of varying amplitude by at least one learning contour identification system, and stored in a data type format readable by at least one learning contour identification system.

8. ) The method of claim 6, wherein a digital data type format comprises a readable transmission and reception of electric signal data types translated into binary data type format where each bit is representative of two distinct amplitudes by at least one learning contour identification system and stored in a data type format readable by at least one learning contour identification system.

9. ) The method of claim 2, wherein transformation of cases into readable case formats by at least one learning contour identification system comprises:

means for the learning contour identification system to transform cases into plurality of contour mappings of contours, with contour pattern metric sets comprising means for the learning contour identification system to process at least one contour mapping of a case'"'"'s contents as a source of data instances for at least one learning contour identification system, and

comprising means of at least one learning contour identification system to maintain the case name labelling for all contours transformed into a plurality of groupings managed and changed by at least one learning contour identification system.

10. ) The method of claim 9, wherein the contour map comprises at least one process whereby contour pattern metric sets of the contour maps of a case, transformed to a plurality of contours by at least one learning contour identification system, do not change a calculated outcome point set of area when the metric container set is deformed within the learning contour identification system.

11. ) The method of claim 9, wherein readable to at least one learning contour identification system comprises at least one of the following:

a translated case data type format represented on a coordinate space dimension greater than zero,

a translated case data type format of a set of coordinate point sets represented on a coordinate system of dimensional space greater than two.

12. ) The method of claim 9, wherein the contour mapping comprises:

means for the learning contour identification system to process data instances into a plurality of contour metrics wherein vital elements of metrics determined by the contour mapping process remains and is stored and unnecessary metrics separated from processing and is stored by at least one learning contour identification system,

wherein the maps of contours are optionally scaled, and wherein distance and direction of the contours are subject to change by the process decision of at least one learning contour identification system, while the relationship between points within the contour map are maintained by at least one learning contour identification system.

13. ) The method of claim 12, wherein the learning contour identification system comprises means for storage of irrelevant information determined by the transformation process, as an additional contour metric, by at least one learning contour identification system.

14. ) The method of claim 1, wherein at least one learning contour identification system comprises:

means for classifying contours of the contour map of contours, and their respective groupings, from contour maps of at least one case, and

means for identifying and labeling contour pattern output from the classification of the contours and their groupings of data items of interest to both users and system processes and their interfaces of the same learning contour identification system.

15. ) The method of claim 1, wherein at least one learning contour identification system comprises means for converting a case into readable data by at least one learning contour identification system.

16. ) The method of claim 1, wherein the learning contour identification system comprises at least one process of combining learning contour identification systems.

17. ) The method of claim 16, wherein the learning contour identification system comprises of subsets of learning contour identification systems.

18. ) The method of claim 1, wherein mapping comprises a machine process of case transformations and data translations.

19. ) The method of claim 18, wherein the four transformations processed in at least one learning contour identification system comprising at least one of:

means for translation,

means for reflection,

means for rotation, and

means for dilation of contours and their metrics within the learning contour identification system.

20. ) The method of claim 18, wherein transforming comprises at least one processing of a super-set of data translations.

21. ) The method of claim 19, wherein case transformation comprises at least one process of creating a correspondence between records and fields of a data source schema to records and fields in a destination schema created by at least one learning contour identification system and stored within the contour identification system.

22. ) The method of claim 20, wherein case translation comprises at least one process of changing the format of a data instance message within the contour identification system.

23. ) The method of claim 1, wherein contour map comprises the contour set where vital information to the user and the learning object identification system of a case remains and unnecessary details determined by same system are removed and remaining data and removed data are stored in memory.

24. ) The method of claim 23, wherein a mapping comprises of at least one:

processing on the contour mapping scaling limits by at least one learning contour identification system, and

processing on the contour distance and direction experience change by at least one learning identification system, while relationship between points describing the contours are maintained by at least one learning contour identification system.

25. ) The method of claim 1, wherein the contour within the contour map of a case is transformed into a plurality of contour metrics binding items of data instances describing patterns of interest found by its learning contour identification system and its system interfaces.

26. ) The method of claim 25, comprising means for each metric to be created for the purpose of determining from the plurality of contour combining any patterns of data instances of interest that can be determined from a collection of training cases of contour metrics, and the testing of other contour metrics, by way of output of training processor machine instructions acting on metrics, for use of pattern identification and labeling of patterns as an objects for learning contour system iterations and user evaluations interfaced to learning contour identification by display and computer higher level language developed applications and input devices.

27. ) The method of claim 26, wherein the contour transformation comprises:

means for creating contour metrics which bound items of data instances of interest to the learning contour identification system as a characterization means which grouped these classifications for decision information processed by at least one learning contour identification system, and,

a storage location to be processor decided upon for memory storage of all metrics of a single contour of a mapping residing between two dynamically adjustable memory addresses of volatile memory used for immediate processing and non-volatile memory used for portability of transformed contours and their metrics.

28. ) The method of claim 27, wherein a non-volatile memory location for each contour of metrics comprises of at least one of the following:

a label location, a location for point-to-point values of the contour,

a filler summation data location,

a location for statistics,

a location for a plurality of mathematical manipulations of statistics, fillers and point-to-point metric values, and,

a location for a plurality for contour combinations of all options.

29. ) The method of claim 28, wherein mathematical calculations are processes of functions of metric values found within a single contour metric and processed within the learning contour identification system.

30. ) The method of claim 28, wherein a non-volatile memory location between two memory address locations of a single contour of the contour mapping is comprised of:

a label identifier metric,

a plurality of contour filler metrics,

a plurality statistic metrics,

a plurality of mathematical processed metrics, and,

a plurality of groupings of contour metrics of similar structure.

31. ) The method of claim 30, wherein the statistic metric comprises the statistical plurality of components of Gaussian Mixture model statistics.

32. ) The method of claim 30, wherein the fillers are unitary weighted summation totals of rows and columns that are placed within the contour boundary the point-to-point metric defines and wherein filler sums are stored as a summation metric of that contour of that contour mapping set from a data case.

33. ) The method of claim 1 wherein training cases comprise a set of more than one case where each case comprises of at least one contour to be transformed into contour metrics.

34. ) The method of claim 1, wherein each metric can have its own label identifier within its block of memory within the two memory address locations defining one contour of each contour mapping of a supplied case to at least one learning contour identification system.

35. ) The method of claim 1, wherein contour metrics of one case can be used in combination of another case transformed into its own contour metrics from its own contour map of contours.

36. ) The method of claim 1 wherein test cases comprise of at least one case.

37. ) The method of claim 1, wherein all processes, interfaces, and learning contour identification systems can be controlled by a higher language machine code instruction set executed to simulate said top level learning control identification systems in a computer hardware system for allowance of making learning contour identification system portable to any computer system and its application software designed to operate as at least one learning contour identification system and to make the contour metrics usable outside the learning contour system that generated contour metrics of contours by way of storage in non-volatile memory through communication channels of the learning contour system.

38. ) The method of claim 1, wherein the contour comprises:

the contour of a case contour mapping of numerical instances resulting from a plurality of mathematical calculations within high level instruction code sets, and

a application modules attached by way of hardware interfaces to at least one learning contour identification system controlled by high level instruction code sets to low-level micro code used by learning contour identification system.

39. ) The method of claim 1, wherein a metric of the contour comprises of a process of storage of metrics representing each contour of a transformed plurality of storage locations of each contour.

40. ) The method of claim 1, wherein the contour metric comprises a minimum of a label metric and a point-to-point representation of a single contour.

41. ) The method of claim 1, wherein a metric of the contour comprises at least one of:

a single contour with its label metric having a dimensional set of numbers,

a single contour with its label metric having a character sequence,

a single contour having a real number metric,

a single contour having a symbol metric,

a single contour metric having a plurality of abscissa stored values and ordinate stored values,

a single contour having a metric that is a dimension of a vector space,

a single contour having a metric that is finite-dimensional,

a single contour metric having a plurality of dynamically changing elements at its memory storage location,

a single contour having a metric that changes dynamically between its location begin and end address in the contour'"'"'s defined memory space,

a single contour whose contour metrics are defined between two memory addresses,

a single contour having its metric storage location sequence order, within its two memory addresses, being a function of the machine code processing sequence of at least one learning contour identification systems,

a single contour whose contour metrics stored within its two memory address contain a plurality of other contour metrics to represent same contour as a grouping of a plurality of contours each of which has a dimension greater than zero,

a single contour having a metric of data point value differences to be used by at least one learning contour system as data point value similarity set to create additions to a contours point-to-point defined by at least one learning contour identification system,

a single contour having a metric of data point values created by processes of a plurality of combinations of mathematical expressions, and,

at least one learning contour identification system'"'"'s single transformed contour, contour metric derived from training case data point value differences, considered as data point likenesses, from a plurality of higher dimensional contour metrics, the learning contour identification system, whose sets are combined to form an appendage to the original contour that are contours of a plurality of contours defined by the sets, used to describe the same single contour.

42. ) The method of claim 1, wherein an contour comprises:

a pattern, not necessarily a user identifiable pattern matching an object known to a noun defined physical classification label, is an identifiable shape structure with contour label metric being determined during the learning identification system process of execution

a pattern of which is identified by classifier means for comparison to training outputs of a

a plurality of past case contour metrics with that of test case contour mapped and transformed set of contour test metrics.

43. ) The method of claim 1, wherein the contour is a graph.

44. ) The method of claim 43, wherein a graph is a diagram representing a system of connections among greater than two things by a number of distinctive point-to-point drawings.

45. ) The method of claim 1, wherein an contour is a network of lines connecting points of a coordinate defined space defined by a case to be transformed into contour metrics.

46. ) The method of claim 1, wherein a case is a set of instances that are transformed by at least one learning contour identification system into a dimensioned image of pixel intensities.

47. ) The method of claim 46, wherein a case is an image of a finite dimensioned set of pixels each having a location on an axis of a coordinate system.

48. ) The method of claim 47, wherein a pixel is a geometrical shape of color.

49. ) The method of claim 48, wherein pixel size is of plurality dimension.

50. ) The method of claim 1, wherein a point-to-point metric is a set of position points on a coordinate axis.

51. ) The method of claim 1, wherein a point is represented by a set of numbers defining its exact location with a coordinate defined space

52. ) The method of claim 1, wherein, the training process to be processed on contour metrics comprises:

a sequence of machine code of at least one learning algorithm and its enhancements that is black box based having decisions and deletions of contour elements of contour metrics transparent to a reporting of a trained output report to a display device and plurality of learning contour identification systems, and

at least one learning algorithm which is rule based, where a set of rules to be used with test cases define the patterns to be identified by at least one learning contour identification system and where rules are instruction sets sent to the classifier for final output of learning contour identification system.

53. ) The method of claim 52, wherein the training process machine code comprises a classification and regression tree learning method whose decision tree rules instruction set is applied to the test cases contour metrics to be classified for final output identification of labeling pattern.

54. ) The method of claim 53, wherein the training process machine code comprises a sequence of machine code implementing a Random Forest model of training on contour metrics.

55. ) The method of claim 53, wherein obtaining the classification data output comprises:

the processing of the contour set metrics using a sequence of instructions language readable and executable by the processor to generate from the respective plurality of contour metrics of the contour representation a labeled pattern output, and,

a system comprising at least one learning contour identification systems, with a plurality of storage devices storing instructions and contour representation metrics that when instructions are executed by said system of learning contour identification systems, an action of said systems perform operations comprising obtaining data of contour representation metrics sets where each metric is a category of metrics representing groups of patterns with a respective multi-dimensional representation, wherein the multi-dimensional representation of the contour metric set is of a pattern that is in a multi-dimensional space.

80. ) The system of claim 1 and claim 56 comprising at least one of the following:

A high-level code set designed to operate a computer system to simulate the learning contour identification system in entirety in a mathematical user interface and coder such that optimizations of combinations of learning identification systems may be found to give increased probability of correctness of pattern labels by both training module and classifier by inspections of confusion matrices.

A high-level code set designed to calculate confusion matrices outputs from training module instruction sets sequences.

A high-level code set designed to feedback implement configurations of learning contour identification systems so that contour data metrics are accurately descriptive of patterns reported by the training module to the classifiers.

A high-level code set designed to use math modules to increase success reported as output of the training module by confusion matrices.

56. ) A system for identifying an contour, comprising:

a training module having a processor and data capture means for pattern recognition,

classifier means having data capture means for assigning an application to recognized patterns,

a controller having a data path interconnecting the training module and the classifier,

memory medium having inputs and outputs connected to the training module and the classifier,

display device adapted to display the output of the system components

application software adapted to communicate with user, and

input device adapted to allow user means for communication with said system

57. ) The system of claim 56, wherein the training module comprises:

at least one learning contour system to capture and store data and convert and store data to a data format to be processable by at least one learning contour system'"'"'s memory system, wherein each training case is transformed into the contour map, whose set of contours are each transformed into contour metrics comprising;

dynamic contour outline data point memory containers, dynamic summation memory containers, dynamic statistics memory containers, a plurality of dynamic training mathematical machine code instruction set outputs, a plurality of dynamic sub-transformed contour metrics of same contour container sets.

a controller that manages memory locations during operations of grouping for contour pattern searching performed within the training module'"'"'s learning firmware instruction sequence of execution using contour metrics.

a processor that executes machine code instruction sets to group all contours of all case data transformed to contour metrics, whose labels are known, to find contour patterns common within a collection of training samples used for training and a within a collection of training samples used for testing out the training modules output of recorded processes used to find said common pattern.

a controller interface to carry out storage to memory the machine instructions necessary to carry out evaluation

a initializer to be used to set all initialization routines necessary for module to operate at beginning and for module to determine a stopping point of execution.

58. ) The system of claim 56, wherein the classification module comprises:

at least one learning contour system to capture and store data and convert and store data to a data format to be processable by at least one learning contour system'"'"'s memory system, wherein each test case is transformed into the contour map, whose set of contours are each transformed into contour metrics comprising;

dynamic contour outline data point memory containers, dynamic summation memory containers, dynamic statistics memory containers, a plurality of dynamic training mathematical machine code instruction set outputs, a plurality of dynamic sub-transformed contour metrics of same contour container sets.

a controller that manages memory locations during operations of extracting from memory the output of the training module and the contour metrics of the test case.

a processor that executes machine code instruction sets of the output of the training module to determine from the extracted contour metrics of the test case the plurality of patterns to be labeled by comparison of training labeled patterns.

a controller interface to carry out storage to memory the machine instructions necessary to carry out evaluation and storage of labeled pattern of test case.

a initializer to be used to set all initialization routines necessary for module to operate at beginning and for module to determine a stopping point of execution.

59. ) The system of claim 56, wherein the controller comprises a microprogrammed control whereby a method of specifying control is one that uses microcode rather than a finite state representation.

60. ) The system claim 59, wherein microcode is the set of micro-instructions that control a processor.

61. ) The system claim 59, wherein microcode comprises means for incorporating microcode dispatch for dynamically scheduled processors to refer to the process of sending an instructions to a queue.

62. ) The system of claim 61, wherein dispatch comprises means for simplifying decoding of instructions to reduce performance impacts of microcode dispatch.

63. ) The system of claim 56, 57, and 59, wherein the controller comprises a hardwired control wherein an implementation of finite state machine control is performed by a collection of programmable logic arrays.

64. ) The system of claim 56, 57, and 59, wherein the processor comprises at least one processor that is a superscalar architecture having advanced pipelining enabling processor to execute more than one instruction per second.

65. ) The system of claim 56, 57, and 59, wherein the microinstruction comprises a representation of low-level instructions, each of which asserts a set of control signals that are active on a given clock cycle as well as provides specifics as to what microinstruction to execute next in the learning contour identification system.

66. ) The system of claim 56, 57, and 59, wherein the micro-operations are RISC-like instructions directly executed by the hardware within the learning contour identification system.

67. ) The system of claim 56, 57, and 59, wherein memory comprises of trace cache as a instruction cache that holds a sequence of instructions with a given starting address in the learning identification system hardware.

68. ) The system of claim 56, wherein control implementation comprises of at least one:

a finite state diagram means for specifying control of each learning contour identification system.

a microprogramming means for specifying control of each learning contour identification system.

69. ) The system of claim 56, wherein a control plan comprises of an instruction set architecture for both the datapath and controller for the processor of the learning contour identification system.

70. ) The system of claim 56, wherein data system captures data by way of at least one of the following:

a wireless communication channel,

a wave guide channel, and

combinations of all.

71. ) The system of claim 56, comprising more than one of the following:

a training module loading contour metrics from a plurality of higher dimensions to create another contour metric,

a classification means for loading contour metrics from a plurality of dimensions to create another contour metric,

a controller that modifies the contour by deleting metrics,

a controller that appends to the contour metric'"'"'s memory location,

a controller whose instructions set groups other systems to form as its output an input to a contour'"'"'s metric storage location,

a controller whose instructions set groups other systems to form as its output an input, and

a controller whose outcome of instructions sets are able to process groups of contours having metrics of dissimilar data types through translations to numerical data types.

72. ) The system of claim 56 wherein, integration of components are designed to exchange data, documents, information, and processor messages between source and targets defined by machine language process codes of the learning contour identification system.

73. ) The system of claim 56 wherein a processor comprises at least one of the following:

a controller to execute machine language codes, and

a controller able to interface to plug-in modules interfaced by datapath and communicated with by microcode instruction sets.

74. ) The system of claim 56, wherein the learning system transforms a data set into a plurality of contours to be translated into contour patterns analyzable and displayable by the training module and its classifier by way of grouping and interfacing additional learning contour identification systems.

75. ) The system of claim 56, wherein classification and display can occur with the training module turned off.

76. ) The system of claim 56, where in learning contour identification system comprises:

a high-level machine code sequence written to control initializations of hardware

a high-level machine code sequence written to accept input from an attached receiving device

a high-level machine code sequence written to accept input from a display device

a high-level machine code written to control the learning contour identification system for automated learning, training, and characterizing of contour mappings and transformations of contour metrics, stopping and starting at user defined points allowing for modifications when necessary.

a high-level machine code written to change training processes without hardware changes to the system so that transformed contour metrics can be added to by user intervention or by plug-in application modules used to enhance selections of groupings of contours.

77. ) The system of claim 76, wherein plug-in application modules may comprise of:

Math routines that send outputs to the machine code instruction set that stores and creates contour metric containers.

User software applications whose outputs provide to the machine code instruction set additions to increase selection accuracy of patterns matching between test and training module outputs the contour metrics of training cases and test cases that have been transformed from their contour maps to individual contours and their metrics

User software applications using outputs of the training module and classifier to assist the user of the displayed output in making probabilistic statements of the pattern reported to be identified.

78. ) The system of claim 77, wherein the display device comprises of means for user viewing of processed events of the learning contour identification system and means for display of learning system requested user prompts requesting interaction with the hardware by way of attached system input device.

79. ) The system of claim 77 and 76, wherein the display device is a plug-in module to be used by at least one learning contour system hardware and interfaced so as to provide autonomous instructions to the learning contour system hardware via and instruction set communication path between output and said learning system.