Ultrasonic Raw-Frame Imaging for AI Liver Disease Quantification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing ultrasonic imaging systems lack the capability to provide quantitative data about liver diseases and integrate this information with ultrasonic images using artificial intelligence models.
Innovation Solution
An ultrasonic imaging system and method that processes ultrasonic raw data to generate multiple raw image frames, including attenuation, scattering, in-phase, quadrature, and frequency spectrum information, and uses an AI model to obtain quantitative data such as fat fraction and liver steatosis severity, displaying this data alongside the images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If ultrasonic raw data is processed to generate multiple raw image frames with different characteristic information and input into an AI model, then quantitative data about liver disease is obtained, but the device complexity increases
Solution Approach 1:
The ultrasonic raw data is segmented into multiple different characteristic information types (amplitude, phase, frequency spectrum, envelope) to create multiple raw image frames. This segmentation allows the AI model to receive diverse input features that collectively provide more comprehensive information for accurate quantitative analysis of liver disease, resolving the contradiction by breaking down complex data into manageable analytical components.
Solution Approach 2:
The patent transforms ultrasonic raw data from a single-dimensional signal into multi-dimensional representation by generating multiple raw image frames with different characteristic information (amplitude, phase, frequency, envelope). This dimensional expansion enables the AI model to analyze liver disease from multiple perspectives simultaneously, improving measurement precision while the systematic approach manages the increased complexity.
2Loss of information
If multiple raw image frames with different characteristic information are generated and processed through an AI model, then comprehensive diagnostic information is obtained, but the processing time increases
Solution Approach 1:
Multiple raw image frames with different characteristic information (amplitude, phase, frequency spectrum, envelope) are generated in advance from the ultrasonic raw data before AI model processing. This preliminary action ensures that all necessary information dimensions are prepared and organized, allowing the AI model to perform comprehensive analysis without time-consuming data preparation during the actual diagnostic process.
3Loss of information
If ultrasonic raw data is processed to generate multiple types of raw image frames, then the information content increases, but the data processing complexity increases
Solution Approach 1:
The complex data processing is segmented into distinct processing streams for different characteristic information types (amplitude, phase, frequency spectrum, envelope). Each stream processes one type of characteristic information independently, which simplifies the processing complexity for each individual stream while collectively providing comprehensive information content when all streams are combined for AI model input.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables the generation and display of ultrasonic images with integrated quantitative data about liver diseases, providing accurate diagnostic information for liver health assessment.
Implementation Method 1
a probe configured to transmit an ultrasonic signal to an object including a liver and receive an ultrasonic echo signal reflected from the object
Implementation Method 2
generate a first raw image frame including attenuation information of the ultrasonic echo signal
Implementation Method 3
generate a second raw image frame including scattering information of the ultrasonic echo signal
Data Source
AI summary
Disclosed is an ultrasonic imaging system including a probe configured to transmit an ultrasonic signal to an object including a liver and receive an ultrasonic echo signal reflected from the object, a display configured to display an ultrasonic image, an input interface configured to obtain user input, memory configured to store an artificial intelligence model, and at least one processor. The at least one processor is configured to obtain ultrasonic raw data by processing the ultrasonic echo signal, generate a plurality of raw image frames including different characteristic information by processing the ultrasonic raw data, obtain quantitative data about a liver disease from the artificial intelligence model by inputting the plurality of raw image frames into the artificial intelligence model, and display the ultrasonic image and the quantitative data about the liver disease together on the display.