Medical Image Segmentation Correction via Atlas Statistical Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for correcting medical image segmentation lack reliability quantification and do not adequately address regions of low reliability, leading to inconsistent segmentation results.
Innovation Solution
A computer-implemented method that determines distributions of corrections by transforming manually generated segmentation corrections into a patient-independent atlas reference system, conducting statistical analysis, and applying the results to improve segmentation accuracy using machine learning algorithms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual corrections are transformed into a patient-independent reference system, then segmentation reliability is improved, but device complexity increases
Solution Approach 1:
The patent introduces an atlas reference system as an intermediary between patient-specific segmentations. Manual corrections from multiple patients are transformed into this common reference frame, allowing statistical analysis and reliability quantification without requiring direct comparison of patient-specific data. This mediator enables reliable segmentation assessment while managing complexity through standardization.
Solution Approach 2:
The patent combines corrections from multiple different patients by transforming them all into a single patient-independent atlas reference system. This merging allows statistical analysis across populations, improving reliability through aggregated data while the reference system manages the complexity of integration.
2Measurement precision
If statistical analysis of corrections is conducted in a patient-independent reference system, then measurement precision of segmentation reliability is improved, but device complexity increases
Solution Approach 1:
The atlas reference system serves as a mediator that enables precise statistical measurement by providing a common framework. Corrections from multiple patients are mapped to this reference, allowing quantitative reliability assessment through statistical analysis while the reference system manages the computational complexity of cross-patient comparisons.
3Manufacturing precision
If machine learning algorithms are trained using transformed statistical data, then segmentation accuracy is improved, but loss of information may occur during transformation
Solution Approach 1:
The patient-independent atlas reference system acts as an intermediary that enables machine learning training through standardized statistical data while preserving essential information. The transformation to reference space allows algorithm training with population-level statistics, and the reference framework maintains sufficient anatomical detail for accurate segmentation without requiring all patient-specific variations.
Data Source
AI summary
Disclosed is a computer-implemented methods of determining distributions of corrections for correcting the segmentation of medical image data, determining corrections for correcting the segmentation of medical image data, training a learning algorithm for determining a segmentation of a digital medical image, and determining a relation between an image representation of the anatomical body part in an individual medical image and a label to be associated with the image representation of the anatomical body part in the individual medical image using the trained machine learning algorithm. The methods encompass reading a plurality of corrections to image segmentations, wherein the corrections themselves may have been manually generated, transforming these corrections into a reference system which is not patient-specific such as an atlas reference system, conducting a statistical analysis of the correction, and applying the re-transformed result of the statistical analysis to patient images. The result of the statistical analysis may also be used to appropriately train a machine learning algorithm for automatic segmentation of patient images. The application of such a trained machine learning algorithm is also part of this disclosure.