ECG Signal Conversion and Integration for Heart State Diagnosis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current electrocardiogram (ECG) analysis methods are limited by the inability to effectively display and analyze the full 10-second raw signal data from 12-lead ECGs, leading to incomplete diagnostic information and the need for invasive procedures, as most ECG reports in PDF format only show 2.5 seconds of data per lead, making it difficult for physicians to accurately diagnose coronary artery diseases without further confirmation.
Innovation Solution
A method and electronic device that convert ECG files from a first file format (e.g., PDF) to a second format (e.g., SVG), perform zero-padding and stacking operations to generate integrated ECG data, and utilize a deep learning model to predict heart state diagnostics, enabling more accurate evaluation of heart conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If ECG reports are displayed in PDF format showing only 2.5 seconds of data per lead, then the report format is simple and easy to generate, but the diagnostic information is incomplete and loses temporal correlation of the full 10-second signal
Solution Approach 1:
The patent converts the ECG data from PDF format into SVG format, creating a digital copy that preserves the complete 10-second signal information. This copying process allows the full temporal data to be retained and processed without losing diagnostic information, while the SVG format enables programmatic access to the complete signal for integration into the diagnostic system.
Solution Approach 2:
The patent transforms the ECG data representation from a visual PDF display format into a structured digital format (SVG) that can be processed computationally. This dimensional change from static visual representation to structured data enables the system to access and integrate the complete temporal signal across multiple leads, preserving information that would be lost in traditional PDF reports.
2Measurement precision
If only 2.5 seconds of ECG data is displayed per lead in the report, then the report generation is straightforward, but the diagnostic accuracy for detecting coronary artery diseases is reduced
Solution Approach 1:
The patent merges the ECG data from all 12 leads into a single integrated representation that preserves the temporal correlation of the complete 10-second signal. By combining the leads in a way that maintains their temporal relationships, the system achieves comprehensive diagnostic information without requiring manual analysis of multiple separate PDF pages, thus improving diagnostic accuracy while maintaining ease of use through automated processing.
Solution Approach 2:
The patent performs preliminary processing of the ECG data by converting it to SVG format and integrating all leads before diagnostic analysis. This preliminary action prepares the complete temporal signal in advance, ensuring that all diagnostic information is available and properly formatted before the actual diagnostic algorithm is applied, thereby improving measurement precision without adding complexity to the user interface.
3Reliability
If invasive cardiac catheterization is performed for confirmation, then diagnostic reliability is improved, but the procedure carries risks and complications
Solution Approach 1:
The patent enables the ECG system to perform self-diagnostic analysis by automatically processing the complete 10-second signal from all 12 leads and generating diagnostic results. This self-service capability allows the system to provide reliable diagnostic information without requiring invasive follow-up procedures, thereby improving diagnostic reliability while eliminating the associated risks and complications of cardiac catheterization.
Solution Approach 2:
The patent implements a feedback mechanism where the complete temporal ECG signal is fully analyzed and the results are fed back to clinicians for decision-making. By providing comprehensive diagnostic information from the full 10-second signal across all leads, the system enables clinicians to make accurate diagnoses without needing invasive confirmation procedures, thus improving reliability while reducing harmful factors.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An electronic device and a method for diagnosing heart state based on electrocardiogram (ECG) are provided. An ECG file (31, F1) is obtained, and the ECG file (31, F1) is in a first file format and includes a plurality of potential traces (ECG_t1,ECG_t2, ECG_t3) of a plurality of leads. The ECG file (31, F1) is converted to a second file format to obtain electrocardiogram data (41, ECG_d1, ECG_d2, ECG_d3) corresponding to multiple leads. Each potential trace (ECG_t1,ECG_t2, ECG_t3) relative to time in the ECG file is converted to the ECG data (41, ECG_d1, ECG_d2, ECG_d3) of each lead. Integrated ECG data (41, ECG_d1, ECG_d2, ECG_d3) associated with the leads is generated based on the ECG data (41, ECG_d1, ECG_d2, ECG_d3) of the plurality of leads through the zero-padding operation and the stacking operation. A diagnostic result (S1) of heart status is generated based on the integrated ECG data (41, ECG_d1, ECG_d2, ECG_d3) and a deep learning model (DM1).