Ultrasound Wound Imaging With 3D Probability Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing ultrasound diagnostic systems struggle to accurately assess the three-dimensional spread and type of decubitus, particularly when the decubitus is large or when bone protrusions hinder probe contact, making it difficult for users with low skill to understand the distribution and type of findings.
Innovation Solution
An ultrasound diagnostic apparatus that calculates presence probabilities for various findings using deep learning and generates three-dimensional probability maps, integrating these probabilities to provide a visual representation of decubitus distribution, along with features to detect failure regions and notify users of redundant scans.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple ultrasound images are captured by moving the probe in multiple directions, then the three-dimensional distribution of decubitus can be observed, but it becomes difficult for users with low skill to accurately understand the type of finding and three-dimensional spread
Solution Approach 1:
The patent replaces manual image interpretation with an automated image recognition system that uses deep learning to automatically identify and classify decubitus findings. The system processes multiple ultrasound images and generates a three-dimensional probability map showing the distribution and characteristics of decubitus, eliminating the need for users to manually analyze multiple images and understand complex three-dimensional spread patterns.
Solution Approach 2:
The patent transforms two-dimensional ultrasound images into a three-dimensional probability map that visually represents the three-dimensional distribution of decubitus. This dimensional transformation allows users to easily understand the spatial spread and depth of decubitus findings without having to mentally reconstruct three-dimensional information from multiple two-dimensional images.
2Measurement precision
If the decubitus is large or bone protrusions prevent normal probe contact, then complete manual examination becomes difficult, but accurate understanding of finding type and three-dimensional spread is needed
Solution Approach 1:
The patent uses an automated image recognition system that processes ultrasound images regardless of probe contact difficulties. The deep learning model automatically identifies decubitus findings even when probe contact is limited by bone protrusions or large decubitus size, generating accurate classification and three-dimensional distribution information without requiring complete manual examination of all areas.
Solution Approach 2:
The patent introduces an image recognition system as an intermediary between the ultrasound imaging process and the diagnostic interpretation. This intermediary automatically analyzes the images, classifies findings, and generates three-dimensional probability maps, bridging the gap between limited probe access and comprehensive diagnostic information.
3Measurement precision
If multiple findings with different symptom progressions are mixed in the decubitus, then accurate identification of finding type becomes difficult, but precise classification is required
Solution Approach 1:
The patent employs a deep learning-based image recognition system that automatically differentiates and classifies multiple types of decubitus findings with different symptom progressions. The system processes ultrasound images and generates a probability map that identifies and categorizes various findings (such as different stages of decubitus, tissue damage types, and other abnormalities), eliminating the need for manual differentiation of complex mixed findings.
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 accurate understanding of decubitus type and distribution by providing clear, three-dimensional visualizations and integrated probability maps, improving diagnostic accuracy and reducing errors due to probe misalignment or incomplete scanning.
Implementation Method 1
an ultrasound probe 21, a position sensor 13 attached to the ultrasound probe 21
Implementation Method 2
capture ultrasound images of a wound portion J
Data Source
Figure 1
Figure 2~3
Figure 4~7
AI summary
An ultrasound diagnostic apparatus (1) includes an ultrasound probe (21), a position sensor (13) that is attached to the ultrasound probe (21) and acquires positional information of the ultrasound probe (21), an image acquisition unit (7) that scans a wound portion of a subject with an ultrasound beam using the ultrasound probe (21) to acquire ultrasound images of a plurality of frames, a probability calculation unit (9) that calculates a presence probability of a finding related to the wound portion from each of the ultrasound images of the plurality of frames, and a probability map generation unit (10) that generates a three-dimensional probability map of the finding on the basis of the positional information of the ultrasound probe (21) acquired by the position sensor (13) and the presence probability calculated by the probability calculation unit (9).