Motion-Compensated Medical Imaging via Edge Entropy Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Medical imaging technologies face challenges in generating clear reconstructions of moving objects, such as a beating heart, due to motion blur, which existing methods struggle to effectively correct without ground-truth images and often require longer scanning times or additional data processing.
Innovation Solution
A method utilizing machine-learning models, specifically neural networks, to generate motion-compensated reconstructions by calculating edge entropy and adjusting the model based on these entropies, allowing for real-time correction of motion blur without the need for ground-truth images and enabling dynamic scanning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If traditional reconstruction methods are used to generate images of moving objects, then the scanning process can be completed with standard procedures, but motion blur occurs and reduces image quality
Solution Approach 1:
The system performs preliminary motion detection during the scanning process by analyzing projections at different angles, then applies motion compensation corrections before final image reconstruction. This preliminary action allows the system to detect and correct motion artifacts before they degrade the final image quality.
Solution Approach 2:
The system implements a feedback mechanism where motion is detected from projection data, compensation parameters are calculated, and corrections are applied iteratively. The edge entropy calculation provides feedback to optimize the compensation process, continuously improving image quality by reducing motion blur through multiple correction passes.
2Manufacturing precision
If ground-truth images are used for motion correction, then motion blur can be reduced, but the complexity of the system increases and additional data processing is required
Solution Approach 1:
The system uses the scan data itself to detect and correct motion without requiring external ground-truth images. By calculating edge entropy from the acquired projections and using this information to drive motion compensation, the system is self-sufficient and does not need additional complex external reference systems or ground-truth data acquisition equipment.
Solution Approach 2:
The system introduces edge entropy as an intermediary metric that bridges the gap between raw projection data and motion correction. Instead of directly comparing with ground-truth images, the edge entropy serves as a mediator that quantifies image quality and guides the compensation process, simplifying the overall system architecture.
3Manufacturing precision
If comprehensive scan data collection is performed to improve reconstruction quality, then more complete information is obtained, but the scanning time increases
Solution Approach 1:
The system applies motion compensation selectively to regions exhibiting motion rather than uniformly processing all scan data. By identifying and correcting only the affected portions of the image using edge entropy analysis, the system achieves high reconstruction quality without the need to process and correct every element of the scan data, thereby reducing overall processing time.
Data Source
AI summary
Devices, systems, and methods obtain scan data that were generated by scanning a scanned region, wherein the scan data include groups of scan data that were captured at respective angles; generate partial reconstructions of at least a part of the scanned region, wherein each partial reconstruction of the partial reconstructions is generated based on a respective one or more groups of the groups of scan data, and wherein a collective scanning range of the respective one or more groups is less than the angular scanning range; input the partial reconstructions into a machine-learning model, which generates one or more motion-compensated reconstructions of the at least part of the scanned region based on the partial reconstructions; calculate a respective edge entropy of each of the one or more motion-compensated reconstructions of the at least part of the scanned region; and adjust the machine-learning model based on the respective edge entropies.