Multi-Viewport Image Registration for Tumor Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current manual measurements for tracking tumors over time in medical imaging are time-consuming and of limited robustness and reproducibility, making it difficult to assess therapy response effectively.
Innovation Solution
A system comprising a first and second viewport for displaying image datasets, with a position indication unit and corresponding position determining unit that uses correspondence information to automatically indicate corresponding positions between images, allowing for easier comparison without deformation, and enabling rotation/translation to facilitate accurate comparison.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual measurements are used to track tumors over time, then flexibility in analysis is maintained, but time consumption increases and measurement precision decreases
Solution Approach 1:
The system performs preliminary automated registration and correspondence determination between images before the user needs to measure tumors. Image datasets are pre-processed to establish spatial relationships, so when the user views the images, the correspondence is already computed and ready for immediate measurement, eliminating time-consuming manual tracking while maintaining precision
Solution Approach 2:
The system creates a virtual copy of the spatial relationship between images through correspondence information. Instead of manually measuring each tumor in each image, the system copies the positional data from one image to its corresponding registered image, allowing rapid measurement transfer that maintains precision without repeating manual measurement processes
2Ease of operation
If elastic registration with warping is applied to display corresponding positions, then visual alignment is improved, but tumor change detection becomes difficult due to artificial deformations
Solution Approach 1:
The system segments the display into multiple viewports, with at least one showing the original unwrapped image and another showing the registered image. This allows the user to view corresponding positions across different viewports without applying visual warping to either image, preserving the true appearance of tumor changes while still enabling easy correspondence identification through the segmented display layout
Solution Approach 2:
The system introduces an intermediary correspondence indication (such as overlay markers or coordinate indicators) that bridges the two images without deforming either one. These indicators show corresponding positions between the original and registered images, allowing visual alignment assistance while keeping the actual tumor morphology unchanged and detectable
3Measurement precision
If corresponding positions are manually identified between registered images, then measurement accuracy is maintained, but productivity decreases due to time-consuming processes
Solution Approach 1:
The system automatically copies positional information from one image to its registered counterpart using pre-computed correspondence data. When a user identifies a tumor in one image, the system copies the corresponding position and measurements to the other image automatically, maintaining measurement accuracy through the established correspondence relationship while dramatically increasing productivity by eliminating manual position-by-position identification
Solution Approach 2:
The system performs self-service by automatically determining corresponding positions and performing measurements without requiring user intervention for each individual position. The correspondence determination unit automatically processes the registered images and identifies matching features, allowing the system to serve itself in the measurement process while maintaining precision through the registered spatial relationships
Data Source
AI summary
A system for displaying a plurality of registered images is disclosed. A first viewport unit displays a representation of a first image dataset in a first viewport. A second viewport unit displays a representation of a second image dataset in a second viewport. A position indication unit enables a user to indicate a position in the first dataset displayed in the first viewport, to obtain a user-indicated position. A corresponding position determining unit determines a position in the second image dataset corresponding to the user-indicated position, to obtain a corresponding position in the second image dataset, based on correspondence information mapping positions in the first image dataset to corresponding positions in the second image dataset. The second viewport unit displays an indication of the corresponding position in the second viewport.