Morphology-Aware Symbolic Representation for Physiological Signal Anomaly Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for anomaly detection in physiological signals, such as ECG signals, suffer from significant information loss during symbolic representation, leading to improper anomaly detection due to user-defined parameters and non-generalized approaches.
Innovation Solution
A system and method that utilize a processor with modules for maxima and minima finding, feature derivation, clustering, symbolic representation, and anomaly detection, which senses physiological signals, derives features from morphology, performs proximity-based clustering, and represents signals symbolically to detect anomalies using a dissimilarity metric.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If piecewise aggregate approximation (PAA) is used to convert time series signal into symbols, then the signal can be represented in symbolic form, but there is huge information loss of the signal during representation
Solution Approach 1:
The patent changes the parameters used for symbolic approximation from fixed user-defined parameters to dynamically derived parameters based on signal morphology features (amplitude differences, sampling point counts). This allows the symbolic representation to adapt to the specific characteristics of each signal, reducing information loss while maintaining the simplicity of symbolic form representation.
Solution Approach 2:
The patent incorporates feedback by using the actual signal characteristics (maxima and minima points, their amplitudes, and spacing) to determine the symbolic approximation parameters. This feedback mechanism ensures that the symbolic representation accurately reflects the underlying signal morphology, thereby minimizing information loss during conversion.
2Ease of manufacture
If user-defined parameters are used for symbolic approximation, then the method can be implemented, but the parameters need to be set and vary from signal to signal, not providing a generalized method
Solution Approach 1:
The patent applies self-service by having the system automatically derive the symbolic approximation parameters from the signal itself rather than requiring external user definition. The algorithm autonomously identifies maxima and minima points, calculates amplitude differences and sampling point counts, and generates appropriate symbolic representations without human intervention, thereby achieving both ease of implementation and generalization across different signals.
Solution Approach 2:
The patent transitions from static user-defined parameters to dynamic signal-derived parameters. The parameters are automatically adjusted based on the specific morphology of each signal (number of maxima/minima, their amplitudes, and temporal spacing), enabling the method to adapt to various signal types without requiring reconfiguration or user input for each case.
3Device complexity
If approximation techniques are used to convert signal into symbols, then the signal can be simplified, but there is huge information loss leading to improper anomaly detection
Solution Approach 1:
The patent changes the approximation parameters to be based on morphologically significant features (amplitude differences between maxima/minima and the number of sampling points between them). These parameters capture the essential characteristics of the signal morphology, allowing for simplified symbolic representation while maintaining sufficient information for reliable anomaly detection.
Solution Approach 2:
The patent extracts only the most critical morphological features from the signal (maxima and minima points, their amplitudes, and spacing) and uses these extracted features to create the symbolic representation. By focusing on these key features rather than attempting to represent the entire signal, the method achieves simplification while preserving the information necessary for accurate anomaly detection.
Data Source
AI summary
The present disclosure addresses the technical problem of information loss while representing a physiological signal in the form of symbols and for recognizing patterns inside the signal. Thus making it difficult to retain or extract any relevant information which can be used to detect anomalies in the signal. A system and method for anomaly detection and discovering pattern in a signal using morphology aware symbolic representation has been provided. The system discovers pattern atoms based on the strictly increasing and strictly decreasing characteristics of the time series physiological signal, and generate symbolic representation in terms of these pattern atoms. Additionally the method possess more generalization capability in terms of granularity. This detects discord/abnormal phenomena with consistency.