ECG Segment Classification Using Adaptive Thresholds and Morphology
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The analysis of large volumes of ECG signals recorded over extended periods is time-consuming due to the inclusion of noise sections, necessitating efficient methods to distinguish analysis-required segments from noise.
Innovation Solution
An ECG data processing method that segments ECG signals, classifies segments as normal, noise, or analysis-required by comparing peak values to adaptive thresholds and morphological patterns, adjusting thresholds based on segment counts, and converting signals to the frequency domain for further analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If all ECG signals recorded over long periods are analyzed, then comprehensive diagnostic coverage is achieved, but analysis time and computational resources increase significantly
Solution Approach 1:
The ECG signal is divided into multiple segments, and each segment is independently analyzed to determine whether it represents a normal section or an analysis-required section. This segmentation allows the system to process only relevant portions of the continuous ECG signal, reducing overall analysis time while maintaining diagnostic coverage.
Solution Approach 2:
The patent extracts and identifies specific segments that require analysis by comparing peak values against threshold values and detecting morphological patterns. By extracting only the analysis-required segments from the continuous ECG signal, the system eliminates unnecessary processing of normal sections, thereby reducing analysis time while preserving comprehensive diagnostic coverage for relevant portions.
2Measurement precision
If threshold values are strictly applied to filter ECG segments, then noise sections are effectively identified, but some analysis-required segments may be incorrectly classified as noise
Solution Approach 1:
The patent applies different classification criteria to different segments based on their local characteristics. By evaluating peak values, morphological patterns, and regularity features specific to each segment, the system adapts its analysis approach locally rather than applying a uniform threshold, thereby reducing misclassification while maintaining effective noise detection.
Solution Approach 2:
The system uses feedback mechanisms where the analysis results from one segment inform the processing of subsequent segments. By continuously adjusting the classification based on detected patterns and regularity assessments, the system refines its threshold application to minimize false positives while maintaining noise detection effectiveness.
3Measurement precision
If adaptive threshold adjustment is implemented based on segment counts, then classification accuracy improves, but processing complexity increases
Solution Approach 1:
The patent implements dynamic threshold adjustment where threshold values are adapted based on the count and characteristics of processed segments. This dynamic approach allows the system to optimize classification precision during processing by adjusting thresholds according to the specific characteristics of the ECG signal being analyzed, rather than using fixed predetermined thresholds.
Solution Approach 2:
The system changes processing parameters (threshold values) based on the state of the input data (segment counts and characteristics). By modifying threshold parameters adaptively during processing, the system achieves higher classification precision while managing complexity through rule-based parameter adjustment rather than complex machine learning models.
Data Source
AI summary
The disclosure relates to an electrocardiogram (ECG) data processing method and a non-transitory recording medium storing an instruction set for executing an ECG data processing method, and more particularly, to a method of extracting an analysis-required section from an ECG signal by considering a data value or a morphological pattern of the ECG signal.