Shared Image Reconstruction Apparatus for MRI Hardware Cost Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current MRI image reconstruction systems require high-performance hardware, which is underutilized, leading to high costs and inefficient resource usage due to the need for powerful CPUs and GPUs to process K-Space data efficiently within a short time frame.
Innovation Solution
An image reconstruction method that calculates the calculation capability requirement of each task and directs tasks with higher requirements to a shared high-performance image reconstruction apparatus, while tasks with lower requirements are processed by local apparatuses, optimizing resource allocation and reducing hardware costs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Power
If high-performance CPU and GPU are used for image reconstruction, then calculation capability is improved, but hardware cost increases
Solution Approach 1:
The patent implements a shared image reconstruction apparatus that serves multiple MRI systems simultaneously. This high-performance apparatus with powerful CPU and GPU resources is not dedicated to a single system but is shared across multiple systems, allowing each system to access high calculation capability without each system needing to purchase its own expensive high-performance hardware. The shared apparatus performs image reconstruction tasks for multiple systems, reducing overall hardware costs while maintaining high calculation capability when needed.
2Power
If high-performance hardware is deployed, then calculation capability is improved, but resource utilization decreases
Solution Approach 1:
The shared image reconstruction apparatus maintains continuous operation by processing reconstruction tasks from multiple different MRI systems in sequence. When one system is performing scanning (and thus generating reconstruction tasks), the apparatus processes those tasks. When that system is idle, the apparatus processes tasks from other systems in the network. This ensures the high-performance hardware is continuously utilized across multiple systems rather than being idle for extended periods, significantly improving overall resource utilization while maintaining high calculation capability.
3Productivity
If image reconstruction is performed quickly, then productivity is improved, but calculation complexity increases
Solution Approach 1:
The patent extracts the complex calculation workload from individual MRI systems and concentrates it in a dedicated shared image reconstruction apparatus. The local MRI systems perform only simple data collection and preliminary processing, then transmit the raw K-space data to the shared apparatus. The shared apparatus contains the powerful CPU and GPU resources needed to handle the complex reconstruction algorithms (such as compressed sensing and parallel imaging techniques) quickly and efficiently. This separation allows fast reconstruction with high calculation complexity to be performed centrally rather than requiring every local system to have complex hardware.
Data Source
AI summary
The present disclosure is directed to image reconstruction techniques used in magnetic resonance imaging. The techniques disclosed include calculating, for each of image reconstruction tasks to be performed, a calculation capability requirement value of the task by a magnetic resonance system, and determining whether the calculation capability requirement value of the task is greater than a predetermined threshold. If so, the task is sent to a shared image reconstruction apparatus, so that the shared image reconstruction apparatus performs the task. Otherwise, the task is sent to a local image reconstruction apparatus, so that the local image reconstruction apparatus performs the task. The techniques described herein facilitate a reduction in hardware cost required for image reconstruction in MRI.