Transitive Deformation Map Generation for Radiotherapy Image Registration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In adaptive/dose-guided radiation therapy, inverse inconsistency and transitivity errors in deformation maps used for image registration complicate the treatment planning process, leading to uncertainty and potential treatment errors due to sequence-dependent measurement and calculation errors.
Innovation Solution
The solution involves generating inverse consistent deformation maps and combining them to create transitive deformation maps, allowing for flexible image deformation sequences without requiring a specific reference image, thereby reducing transformation artifacts and ensuring consistent results regardless of the deformation pathway.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If deformation maps are generated by comparing images taken at different times during treatment, then image registration can be performed to track tissue changes, but inverse inconsistency and transitivity errors occur making the transformation results highly dependent on the sequence of warping operations
Solution Approach 1:
The patent applies preliminary action by pre-computing and storing deformation maps for all possible image pairs before actual registration is needed. When registration is required, the system simply retrieves the pre-computed deformation map corresponding to the desired image pair, eliminating the need to perform warping operations in sequence and thus avoiding inverse inconsistency and transitivity errors.
2Reliability
If a predetermined reference image is used through which all transformations must be undertaken, then transitivity errors are reduced, but the system loses flexibility in choosing deformation pathways and requires a specific reference image
Solution Approach 1:
The patent applies universality by creating a comprehensive deformation map library that contains pre-computed deformation maps for all possible image pairs. This universal library allows any image to serve as a reference for any other image without requiring a predetermined reference image, thus maintaining both reliability (no transitivity errors) and flexibility (any image can be the reference).
Data Source
AI summary
Deformation maps (e.g. deformation vector fields) used for correcting image-type data used in the treatment of patients in radiotherapy may be processed to eliminate inverse inconsistency and transitivity type errors which produce different results depending on the order or path of the calculation of deformation. The correction permits registration of a treatment plan with the changing patient image and accumulation of dose to a common reference frame without transformation dependent artifacts.