Echocardiogram AI Frequency Analysis for Heart Structure Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing echocardiogram analysis using deep learning methods fails to consider the frequency variation of heart rate, limiting the accuracy of heart structure and shape identification.
Innovation Solution
An echocardiogram artificial intelligence processing system that splits echocardiogram videos into frames, extracts gray values at specific points, converts them into frequency charts, averages these charts, and uses a classifying model to distinguish between normal and abnormal heart conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If static images are used to identify heart structure and shape, then the identification process is simple, but the accuracy is limited because heart rate frequency variation is not considered
Solution Approach 1:
The echocardiogram video is segmented into multiple frames, and each frame is further divided into multiple position points. This segmentation allows the system to analyze temporal variations in heart structure while maintaining manageable processing complexity through systematic breakdown of the video data.
Solution Approach 2:
The patent transforms the problem from static 2D image analysis to dynamic time-frequency analysis by converting gray value variations into frequency charts using Fourier transform. This adds a temporal frequency dimension to the analysis, enabling detection of heart rate variations while maintaining computational feasibility.
2Extent of automation
If deep learning models are trained with relative samples to automatically mark heart size and define structure, then automation is improved, but the model fails to account for heart rate frequency variation
Solution Approach 1:
The patent introduces frequency charts as an intermediary representation between the raw video frames and the final classification. The Fourier transform converts temporal gray value variations into frequency domain representations, which then serve as enhanced input features for the deep learning model, enabling it to automatically detect heart rate variations without manual intervention.
Solution Approach 2:
The system automatically extracts and utilizes frequency parameters by applying Fourier transform to convert temporal variations into frequency domain characteristics. This parameter transformation enables the deep learning model to inherently consider heart rate frequency variations, improving diagnostic accuracy while maintaining full automation.
3Measurement precision
If echocardiogram videos are analyzed in full dynamic form, then heart rate frequency information is preserved, but processing complexity and time increase
Solution Approach 1:
The patent extracts only the essential frequency information needed for diagnosis by applying Fourier transform to convert temporal variations into frequency domain representations. This extraction isolates the critical heart rate frequency characteristics from the full video data, enabling efficient processing while preserving diagnostically relevant information.
Solution Approach 2:
By transforming the problem from temporal domain to frequency domain using Fourier transform, the system efficiently captures heart rate frequency information in a compressed representation. This parameter transformation reduces processing requirements compared to analyzing full dynamic video while maintaining retention of critical frequency characteristics.
Data Source
AI summary
An echocardiogram artificial intelligence processing system includes an ultrasound detecting device and a processor. The ultrasound detecting device obtains an echocardiogram video. A splitting module splits the echocardiogram video into a plurality of frames. Each frame includes a plurality of position points. A gray value obtaining module is to obtain a gray value of each of the position points of each of the frames, and to collect the gray value corresponding to an identical one of the position points of each frame to form a gray value variation chart corresponding to the identical one of the position points. A frequency chart converting module converts the gray value variation chart corresponding to each position point into a frequency chart. A frequency averaging module obtains an echocardiogram variation feature chart. An artificial intelligence classifying module is signally connected to the frequency averaging module.