Probability Distribution Display for Lesion Matching in Tomographic Imaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In medical imaging, accurately matching lesions across different tomographic images captured by various modalities is challenging due to differences in modalities and subject deformation, leading to significant positional errors and inefficient search processes for corresponding lesions.
Innovation Solution
An information processing system that calculates and displays an existence probability distribution on ultrasonic tomographic images, indicating the likely range of a corresponding lesion's location, thereby guiding the user to efficiently identify the lesion while considering positional estimation errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual matching is performed by an operator while viewing corresponding images, then the accuracy in identifying corresponding lesions can be improved through visual inspection, but the search time and workload increase significantly
Solution Approach 1:
The patent replaces the manual visual inspection process (mechanical/operator-based system) with an automated image processing system that calculates existence probability distributions. The system automatically processes tomographic images to determine the likelihood of lesion correspondence, eliminating the need for operators to manually search through images while maintaining or improving accuracy.
Solution Approach 2:
The patent introduces an intermediary computational layer that processes the relationship between tomographic images. Instead of directly comparing images visually, the system uses probability distribution calculations as an intermediary to guide the identification of corresponding lesions, making the matching process more efficient and accurate.
2Reliability
If the entire ultrasonic tomographic image is searched to identify the corresponding lesion, then the possibility of finding the lesion is maximized, but the search efficiency decreases significantly
Solution Approach 1:
The patent segments the search space by dividing the ultrasonic tomographic image into regions with different existence probabilities. Instead of treating the entire image as a uniform search area, the system identifies and highlights specific regions where lesions are more likely to exist, allowing operators to focus their search on high-probability areas rather than scanning the entire image.
Solution Approach 2:
The patent applies local quality by providing different levels of information guidance in different regions of the image. High-probability regions receive enhanced visual guidance or annotations, while low-probability regions receive minimal or no guidance, optimizing the search process by directing attention to where it is most needed.
3Device complexity
If a fixed boundary line is drawn on the ultrasonic tomographic image to indicate the lesion position, then the display is simple and clear, but it does not account for positional deviation and measurement errors
Solution Approach 1:
The patent transitions from a static fixed boundary line to a dynamic probability distribution representation. Instead of displaying a single deterministic boundary, the system dynamically adjusts the display to show probability distributions that reflect measurement uncertainties and positional deviations, allowing the visualization to adapt to the actual confidence levels in the lesion localization.
Solution Approach 2:
The patent incorporates beforehand cushioning by pre-calculating and displaying probability distributions that account for potential measurement errors and positional deviations. This allows the system to prepare for uncertainties in advance, providing a buffer zone or confidence region around the estimated lesion position rather than relying on a single precise boundary that may be inaccurate.
Data Source
AI summary
An information processing apparatus includes a region acquisition unit configured to obtain a specific region of a subject, a tomographic image acquisition unit configured to obtain a tomographic image of the subject, and a display control unit configured to cause a display unit to display a region indicating probability of existence of the specific region in the obtained tomographic image.