Method and apparatus for pattern processing

• US 8,463,042 B2
• Filed: 05/22/2009
• Issued: 06/11/2013
• Est. Priority Date: 05/27/2008
• Status: Expired due to Fees

• Current Assignee: Eberhard Falk
• Original Assignee: Eberhard Falk
• Inventors: Falk, Eberhard
• Primary Examiner(s): DULANEY, KATHLEEN YUAN
• Application Number:
  US12/471,047
• Publication Number:
  US 20100067795A1
• Time in Patent Office: 1,481 Days
• Field of Search: 382/182, 382/161, 382/190, 382/209, 382/276, 706/20
• US Class Current: 382/182
• CPC Class Codes: G06F 18/28 Determining representative reference patterns, e.g. by averaging or distorting; Generating dictionaries
• CPC Class Codes: G06V 30/10 Character recognition

Abstract

An apparatus for pattern processing exhibits a discretizing device for discretizing an input pattern, a device for generating a number n of discrete variants of the quantized input pattern in accordance with established rules, a number n of input stages (50) for generating, for each input-pattern variant, an assigned output symbol from a set of symbols, and a selection unit (60) for selecting a symbol by way of selected symbol relating to the input pattern from the n generated output symbols in accordance with an established selection rule. The apparatus according to the invention and the corresponding process according to the invention enable a faster, more precise and more flexible recognition of patterns, in which connection it may be a question of spatial image patterns, temporally variable signal patterns and other input patterns.

Claims

1. Process for pattern processing, exhibiting the following steps:

a) discretizing of an input pattern,

b) generation of a number n of discrete variants of the discretized input pattern in accordance with established rules,

c) input of each of the n discrete input-pattern variants into a memory element and, in each memory element, mapping of the entered input-pattern variant to a respectively assigned output symbol from a finite set of discrete symbols in accordance with an associative assignment rule, and

d) selection of one of the output symbols by way of selected symbol relating to the input pattern from n generated output symbols in accordance with an established selection rule,

wherein the associative assignment rule in step c) is learned by feedback of the selected symbol as a reference symbol at a symbol input (set) of the memory element;

in which

the pattern processing is executed as a multi-stage process, wherein selected symbols of a first process stage serve as input patterns of a second process stage, and in which

the generation of a number n of discrete variants of the discretized input pattern in process step b) of the second process stage is effected by mixing output symbols of the first stage.

2. Process according to claim 1, wherein the input pattern is a spatially variable input image.

3. Process according to claim 2, wherein the discrete variants of the discretized input pattern in process step b) are generated in each instance by one or more image-processing operations selected from the group consisting of displacements, rotations, scalings and threshold-value operations on the input pattern.

4. Process according to claim 1, wherein the input pattern is a temporally variable input signal.

5. Process according to claim 1, wherein the finite set of discrete symbols in process step c) includes a symbol with the meaning “

indeterminate”

.

6. Process according to claim 1, wherein the established selection rule in process step d) includes a majority decision.

7. Process according to claim 6, wherein in a case of the majority decision, a simple majority of the symbols suffices for determining the selected symbol.

8. Process according to claim 6, wherein in a case of the majority decision a number of other assigned symbols (votes against) are registered.

9. Process according to claim 1, wherein the number n of discrete input-pattern variants is amended dynamically.

10. Process according to claim 1, wherein the number n of discrete variants of the discretized input pattern is amended during the process.

17. Computer program embodied on a non-transitory computer-readable medium, which, when running on a computer system, executes the process according to claim 1.

18. Non-transitory computer-readable medium, on which the computer program according to claim 17 is stored.

11. Apparatus for pattern processing, exhibiting:

a discretizing device for discretizing an input pattern;

a device for generating a number n of discrete variants of the quantized input pattern in accordance with established rules;

a number n of memory elements which have been set up to map each of the n discrete input-pattern variants onto a respectively assigned output symbol from a finite set of discrete symbols in accordance with an associative assignment rule; and

a selection unit for selecting one of the output symbols by way of selected symbol relating to the input pattern from n generated output symbols in accordance with an established selection rule;

wherein the memory elements and the selection unit are connected to one another in such a way that the associative assignment rule which is used in at least one of the connected memory elements is learned by feedback of the selected symbol as a reference symbol at a symbol input (set) of the connected memory elements; and

wherein the memory elements each exhibit a pattern input, a reference-symbol input, an update-activation input and a symbol output.

12. Apparatus according to claim 11, wherein the memory elements take the form of associative memories.

13. Apparatus according to claim 11, wherein the selection unit exhibits a number n of symbol inputs, a symbol output, an output for outputting a signal (majority) that specifies a number of the selected symbol and an output for outputting a signal (veto) that specifies a number of other input symbols.

14. Apparatus according to claim 13, wherein the selection unit additionally exhibits a reference-symbol input and an update-activation input.

15. Apparatus according to claim 11, which takes a form of a multi-stage pattern-processing apparatus, wherein selected symbols of a first processing stage serve as input patterns of a second processing stage.

16. Apparatus according to claim 15, wherein mixers are provided for mixing output symbols.