Adaptive Medical Image Reconstruction for ROI
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current medical diagnostic image reconstruction techniques can be time-consuming, particularly for high-fidelity images, leading to delays in urgent care environments and inefficiencies in identifying clinical regions of interest, and they do not adaptively reconstruct images based on user-defined areas of interest.
Innovation Solution
The system stores non-reconstructed image data and performs adaptive reconstruction only when a user selects a region of interest, using reconstruction engines to generate higher-quality images on demand, allowing for progressive and iterative reconstruction of selected areas while ignoring or minimizing updates to other regions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If full image reconstruction is performed for the entire dataset, then high-fidelity diagnostic images are produced, but reconstruction time increases significantly causing delays in urgent care environments
Solution Approach 1:
The patent divides the entire image dataset into multiple regions of interest (ROIs) based on clinical relevance. Instead of reconstructing the full dataset, the system identifies and reconstructs only specific segmented regions that contain diagnostically important information, thereby reducing reconstruction time while maintaining image quality for critical areas.
Solution Approach 2:
The patent applies different reconstruction quality levels to different regions of the image dataset. High-fidelity reconstruction is applied locally to identified regions of interest, while other areas receive lower priority or no reconstruction. This local quality approach ensures diagnostic accuracy where needed without the time cost of full-dataset high-fidelity reconstruction.
2Productivity
If adaptive reconstruction based on user-defined regions of interest is implemented, then reconstruction efficiency improves and time is reduced, but system complexity increases due to additional processing requirements
Solution Approach 1:
The patent performs preliminary identification of regions of interest using low-cost methods such as analyzing projection data metadata, detecting high-attenuation structures, or using quick preview reconstructions. These preliminary actions prepare the data structure and identify candidate ROIs before the actual adaptive reconstruction process, reducing the complexity of the main reconstruction algorithm by pre-organizing the work.
Solution Approach 2:
The patent introduces an intermediary layer between data acquisition and final reconstruction that handles ROI identification and selection. This intermediary component processes the raw data to extract meaningful regions, then passes only those regions to the reconstruction engine. This mediation simplifies the overall system architecture by separating concerns and allowing each component to specialize.
Data Source
AI summary
Example methods, apparatus and articles of manufacture to adaptively reconstruct medical diagnostic images are disclosed. A disclosed example method includes storing non-reconstructed image data captured by a medical image acquisition system, receiving a parameter representing a region of interest from a diagnostic imaging workstation, and communicating a portion of the non-reconstructed image data associated with the region of interest to the diagnostic imaging workstation in response to receiving the parameter, wherein the portion of the non-reconstructed image data is processed by the diagnostic imaging workstation to form a medical diagnostic image.