Machine Learning Reconstruction Duration Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In medical imaging, particularly in magnetic resonance imaging, the reconstruction time is often lengthy, and existing methods lack a systematic approach to predict and optimize reconstruction durations, leading to inefficiencies and user frustration due to the complex interplay of hardware, software, and protocol parameters.
Innovation Solution
A computer-implemented method using a trained advance calculation function based on machine learning to predict reconstruction durations by compiling input datasets of protocol and hardware/software parameters, allowing for informed planning and optimization of imaging processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Duration of action of moving object
If recording time is shortened using novel reconstruction methods (SMS, CS, CAIPIRINHA, Deep Learning), then recording time is reduced, but reconstruction time increases significantly
Solution Approach 1:
The system performs preliminary calculation of reconstruction durations by compiling input datasets containing protocol parameters, hardware parameters, and software parameters before the actual imaging process. This allows users to plan imaging processes in advance with knowledge of expected reconstruction times, enabling optimization of protocol selection and ordering without requiring trial-and-error approaches.
2Loss of information
If manual decision trees are used to estimate reconstruction durations, then some planning information is available, but the process is extremely effortful and error-prone due to complex dependencies
Solution Approach 1:
The patent replaces manual decision tree approaches with an automated machine learning-based calculation function. The system automatically compiles input datasets from protocol parameters, hardware parameters, and software parameters, and uses a trained machine learning model to predict reconstruction durations. This eliminates the need for manual navigation through complex decision trees while handling the intricate dependencies between multiple parameters automatically and accurately.
3Duration of action of stationary object
If users adjust protocol parameters to optimize reconstruction time, then reconstruction duration can be reduced, but users lack systematic guidance and must rely on trial-and-error methods
Solution Approach 1:
The system provides feedback to users by calculating and displaying predicted reconstruction durations based on selected protocol parameters, hardware configuration, and software version. This feedback mechanism enables users to understand the impact of different protocol settings on reconstruction time and make informed decisions about parameter optimization without needing to perform trial-and-error recordings. Users can plan optimal imaging protocols in advance based on this predictive information.
Data Source
AI summary
The disclosure relates to techniques for operating an imaging facility for preparing an imaging process. For each imaging process, at least one image dataset is reconstructed in a reconstruction step from raw data recorded in accordance with at least one recording protocol using a reconstruction facility with reconstruction software. For advance calculation of a duration for the reconstruction step, an input dataset comprising at least one protocol parameter of the recording protocol influencing the duration of the reconstruction step and at least one hardware parameter describing the hardware of the reconstruction facility and/or at least one software parameter describing the reconstruction software is compiled, and the duration is ascertained from the input dataset by way of a trained advance calculation function, which is trained by machine learning.