Medical Image Deformation Estimation Using Precomputed Basis Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current medical image processing techniques face challenges in quickly and easily aligning images captured from different body positions, such as MRI and ultrasonic images, due to varying deformation states, which requires time-consuming deformation simulation and complex finite element model generation.
Innovation Solution
A processing apparatus that obtains shape data from multiple subjects in different states, generates basis data for deformation, and estimates deformation for a target subject using statistical models, enabling rapid alignment of medical images across different positions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If deformation simulation is executed using finite element model, then deformation estimation precision is improved, but processing time increases significantly
Solution Approach 1:
The patent pre-calculates and stores deformation data for multiple body positions (prone, supine, lateral) using finite element models during an offline phase. During the online clinical phase, these pre-computed deformation fields are simply retrieved and applied, eliminating the need for real-time deformation simulation while maintaining high precision.
Solution Approach 2:
The patent creates a library of pre-computed deformation fields that copy the results of complex finite element simulations. Instead of executing the full simulation again during clinical use, the system uses these pre-generated deformation field copies, which capture the essential deformation characteristics without requiring computational resources at the time of actual processing.
2Measurement precision
If finite element model is generated for each target case, then deformation estimation accuracy is improved, but operational complexity increases
Solution Approach 1:
The patent develops a universal deformation field library that serves multiple patients and body positions. Instead of creating patient-specific models for each case, the system uses a single comprehensive library containing deformation fields for various body positions that can be applied to any patient, significantly simplifying the operational workflow.
Solution Approach 2:
The patent parameterizes deformation fields by body position (prone, supine, lateral) rather than by patient anatomy. This parameterization approach allows the system to select the appropriate deformation field based on the imaging position without requiring custom model generation for each patient, reducing operational complexity while maintaining accuracy.
Data Source
AI summary
An image processing apparatus obtains, for each of a plurality of subjects, a data set including first shape data which indicates a shape of a subject measured in association with the subject in a first state, and second shape data which indicates a shape of the subject measured in association with the subject in a second state, obtains basis data required to express a deformation from the first state to the second state, based on the data sets for the plurality of subjects, and estimates, based on the generated basis data and data indicating a shape of a target subject measured in association with the target subject in the first state, a deformation from the first state to the second state in association with the target subject.