Patient Image Analysis Algorithm Versioning for Output Comparability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for analyzing patient image data sets lack comparability due to changes in analysis algorithms and workstation environments over time, leading to inconsistent outputs.
Innovation Solution
A method involving the use of containers or virtual machines to store and manage analysis algorithms, ensuring that the same algorithm is used for both initial and subsequent image data sets, even if the workstation environment changes, by storing algorithms in a server database and transferring the appropriate container or virtual machine to the workstation for analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If analysis algorithms are updated over time to improve analysis capabilities, then analysis quality and accuracy are improved, but output comparability between different image data sets deteriorates
Solution Approach 1:
The system stores the specific version identifier of the analysis algorithm that was used to process each image data set at the time of processing. This preliminary recording of algorithm version information enables later retrieval and reapplication of the exact same algorithm version, ensuring that outputs remain comparable across different time points even as newer algorithm versions become available
Solution Approach 2:
When comparability is required, the system copies the exact algorithm version that was previously used and applies it to new or reprocessed image data sets. This copying of the specific algorithm version (rather than using the latest available version) ensures that analysis outputs remain consistent and comparable across different processing instances
2Stability of the object's composition
If the same analysis algorithm is used for multiple image data sets to ensure comparability, then output comparability is maintained, but the ability to benefit from updated algorithms is lost
Solution Approach 1:
The system dynamically selects which algorithm version to apply based on the processing requirements. When comparability is needed, it selects the previously used version; when processing new data where latest capabilities are desired, it can select updated algorithm versions. This dynamic selection mechanism allows the system to adapt between maintaining stability and incorporating improvements based on the specific processing context
3Productivity
If analysis algorithms change over time, then analysis capabilities are improved, but consistency of analysis results across different time points deteriorates
Solution Approach 1:
The system introduces an intermediary mechanism (version identification and storage system) that mediates between the evolving analysis algorithms and the requirement for consistent results. By recording and managing algorithm version information as an intermediary layer, the system can retrieve and reapply specific historical versions when consistency is required, while still allowing progression to newer versions when capability improvement is the priority
Data Source
AI summary
The present invention suggests a method for analysing image data (21, 22) of a patient by means of a server (10) having a database (15) for analysis algorithms (11, 12) comprising the steps: - providing a first image data set (21) of a patient, - performing an analysis of the first image data set (21) by using a first analysis algorithm (11), - adding a second analysis algorithm (12) to the database, - providing a second image data set (22), - assigning the second image data set (22) to the patient and - selecting the first analysis algorithm (21) from the database (15) for performing the analysis of the second image data set (22).