Signal and pattern detection or classification by estimation of continuous dynamical models

• US 6,564,176 B2
• Filed: 08/17/2001
• Issued: 05/13/2003
• Est. Priority Date: 07/02/1997
• Status: Expired due to Term

• Current Assignee: Adobe Systems Incorporated (Adobe Inc.)
• Original Assignee: Nonlinear Solutions, Inc.
• Inventors: Kadtke, James B., Kremliovsky, Michael N.
• Primary Examiner(s): ASSOUAD, PATRICK J
• Application Number:
  US09/932,450
• Publication Number:
  US 20020133317A1
• Time in Patent Office: 634 Days
• Field of Search: 702/189, 702/196, 703/2, 704/243, 700/52
• US Class Current: 702/189
• CPC Class Codes: G06F 18/21 Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation

Abstract

A signal detection and classification technique that provides robust decision criteria for a wide range of parameters and signals strongly in the presence of noise and interfering signals. The techniques uses dynamical filters and classifiers optimized for a particular category of signals of interest. The dynamical filters and classifiers can be implemented using models based on delayed differential equations.

Claims

1. A method for detecting and classifying signals, comprising:

acquiring a data signal from a dynamical system;

pre-processing the data signal;

estimating the preprocessed signal'"'"'s derivative to generate a derivative vector;

applying an expansion function to the derivative vector to generate estimated coefficients; and

processing the estimated coefficients to classify the data signal.

2. A method for detecting and classifying signals as defined in claim 1, wherein pre-processing the data signal comprises normalizing the data signal.

3. A method for detecting and classifying signals as defined in claim 1, wherein pre-processing the data signal comprises filtering the data signal.

4. A method for detecting and classifying signals as defined in claim 1, wherein pre-processing the data signal comprises smoothing the data signal.

5. A method for detecting and classifying signals as defined in claim 1, wherein pre-processing the data signal comprises transforming the data signal.

6. A method for detecting and classifying signals as defined in claim 1, wherein processing the estimated coefficients comprises applying a density estimator to the estimated coefficients to classify the data signal.

7. A method for detecting and classifying signals, comprising:

acquiring a data signal from a dynamical system;

pre-processing the data signal;

estimating the preprocessed signal'"'"'s derivative to generate a derivative vector; and

applying an expansion function to the derivative vector to generate estimated coefficients by performing a polynomial expansion on the derivative vector to generate a design matrix, and decomposing the design matrix to generate a feature vector having the estimated coefficients.

8. A method for detecting and classifying signals as defined in claim 7, wherein the polynomial expansion is a simple polynomial expansion.

9. A method for detecting and classifying signals as defined in claim 7, wherein the polynomial expansion is a rational polynomial expansion comprising a ratio of two polynomial expansions.

10. A method for detecting and classifying signals, comprising:

acquiring a data signal from a dynamical system;

pre-processing the data signal;

estimating the preprocessed signal'"'"'s derivative to generate a derivative vector; and

applying an expansion function to the derivative vector to generate estimated coefficients by performing a orthogonal function expansion on the derivative vector to generate a design matrix, and decomposing the design matrix to generate a feature vector having the estimated coefficients.

11. A method for detecting and classifying signals, comprising:

acquiring a data signal from a dynamical system, wherein the data signal comprises at least one two-dimensional image;

pre-processing the data signal;

estimating the preprocessed signal'"'"'s derivative to generate a derivative vector; and

applying an expansion function to the derivative vector to generate estimated coefficients.