ECG Signal Quality Evaluation via Multi-Scale Densely Connected Network
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current ECG signal quality evaluation methods struggle to accurately assess the quality of signal fragments and extract local signal features, and they often require a large number of parameters and computations, making them unsuitable for real-time evaluation on portable or wearable devices.
Innovation Solution
A method using a multi-scale convolutional and densely connected network for ECG signal quality evaluation, which involves preprocessing ECG signals to remove noise, segmenting them into fragments, and using a trained AlexNet model to correct labels before inputting them into an improved lightweight densely connected quality classification model for evaluation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If current neural network models are used for quality evaluation, then evaluation accuracy can be achieved, but the number of parameters and computation amount become large
Solution Approach 1:
The patent segments the ECG signal into multiple signal fragments and processes each fragment independently through the neural network. This segmentation approach allows the model to focus on local features within each fragment, reducing the overall computational burden while maintaining evaluation accuracy for each segment. The model evaluates quality at the fragment level rather than processing entire long signals at once.
Solution Approach 2:
The patent implements local quality evaluation by analyzing specific signal fragments rather than treating the entire signal uniformly. The neural network is designed to extract local features from each fragment and assign quality labels independently. This local approach allows the system to identify and evaluate problematic regions without being constrained by the quality of other segments, reducing the need for complex global analysis.
2Measurement precision
If entire ECG signals are evaluated, then overall quality assessment is achieved, but local signal features and fragment quality cannot be accurately evaluated
Solution Approach 1:
The patent divides the continuous ECG signal into discrete fragments of fixed duration. Each fragment is processed independently by the neural network, which extracts local features and assigns quality labels specific to that fragment. This segmentation preserves local quality information that would be lost in global evaluation, as each fragment's characteristics are analyzed in isolation.
Solution Approach 2:
The patent introduces signal fragments as an intermediary between the raw ECG signal and the final quality evaluation. These fragments serve as intermediate units that capture local signal characteristics while being manageable in size for neural network processing. The fragment-level evaluation acts as a mediator that preserves local information before aggregating results for overall signal assessment.
3Measurement precision
If complex neural network models are used, then evaluation accuracy improves, but real-time processing capability is compromised
Solution Approach 1:
The patent processes ECG signals in segmented fragments rather than as complete long-duration signals. This segmentation reduces the computational load per processing unit, enabling faster inference times that meet real-time requirements. The neural network processes each short fragment independently and rapidly, allowing for real-time quality monitoring without sacrificing accuracy on each local segment.
Data Source
AI summary
An electrocardiograph (ECG) signal quality evaluation method based on a multi-scale convolutional and densely connected network is provided. Firstly, an original ECG signal is preprocessed to remove a baseline drift and power line interference. Then, based on a consistency principle of a label determining result and a principle of setting a confidence coefficient, an AlexNet model is trained to mutually correct incorrect labels in a dataset to obtain a final ECG signal fragment for quality classification. Finally, the signal fragment is input into an improved lightweight densely connected quality classification model to classify quality of the ECG signal fragment.