Medical Image Registration by Functional Segment Alignment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing functional diagnostic methods face challenges in accurately aligning functional images of a human body, leading to low diagnostic accuracy due to variations in anatomical structures within organs, especially when the function to be observed is localized or differs by site.
Innovation Solution
A medical image processing apparatus and method that includes obtaining morphological and functional images, segmenting them based on a function of interest, and performing registration processes to align segmental images, ensuring precise alignment of anatomical and functional units.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If functional images are aligned based on outer shape or organ boundaries, then alignment process is simple, but alignment accuracy at anatomical structure level inside the organ deteriorates
Solution Approach 1:
The patent segments the organ into multiple anatomical structures based on functional information. By dividing the organ into functional segments (e.g., functional regions with different metabolic activities), the system enables precise alignment at the anatomical structure level rather than treating the entire organ as a single unit. This segmentation allows accurate registration of functional images with morphological images at the level of individual anatomical structures within the organ.
Solution Approach 2:
The patent applies different alignment strategies to different regions of the organ based on their functional characteristics. By identifying regions with distinctive functional features (such as areas with different glucose metabolism patterns), the system performs localized alignment operations that optimize accuracy for each specific anatomical region rather than applying a uniform alignment approach to the entire organ.
2Productivity
If functional images are aligned using conventional methods, then processing time is short, but diagnostic accuracy deteriorates due to inability to properly compare localized functions
Solution Approach 1:
The patent performs preliminary segmentation of the organ into functional structures before the alignment process. By pre-dividing the organ into functional segments and preparing functional reference data, the system enables rapid yet accurate alignment operations. This preliminary organization of functional information allows the alignment algorithm to work more efficiently while maintaining high diagnostic accuracy for localized functional comparisons.
Solution Approach 2:
The patent introduces functional reference data as an intermediary between the functional image and the morphological image. This functional reference serves as a mediator that contains information about the anatomical structures' functional characteristics, enabling the system to accurately register functional images with morphological images by matching both anatomical and functional features rather than relying solely on one type of reference.
Data Source
AI summary
A medical image processing apparatus includes processing circuitry. The processing circuitry obtains a morphological image and a functional image including at least a predetermined anatomical structure of a test subject; obtains reference morphological data and reference functional data corresponding to the morphological image and the functional image, respectively; individually segments the morphological image and the reference morphological data according to a function of interest appearing on the functional image; segments the functional image with reference to the segmental morphological images, and segments the reference functional data with reference to the segmental reference morphological data; performs a first registration process for aligning the segmental morphological images and the segmental reference morphological data with each other; and performs a second registration process for aligning the segmental functional images and the segmental reference functional data with each other according to a result of the first registration process.