Attenuation Coefficient Image Generation Using Intermediate Tissue Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for generating attenuation coefficient images in nuclear medicine diagnostics without CT or MR imaging may produce values outside appropriate ranges, making it difficult to ensure accurate attenuation coefficient values.
Innovation Solution
An attenuation coefficient image generation method that involves generating an input image from measurement data, an intermediate image related to tissue areas, and using known attenuation coefficients to produce a reliable attenuation coefficient image, along with a trained model generation method using pseudo-data for training without requiring extensive clinical images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If an attenuation coefficient image is generated from PET data using a machine learning model without CT or MR imaging, then the imaging process is simplified and time is reduced, but the attenuation coefficient values may fall outside appropriate ranges reducing reliability
Solution Approach 1:
The patent introduces an intermediate image as a mediator between the input PET image and the final attenuation coefficient image. This intermediate image contains tissue area information that guides the generation process, ensuring that the machine learning model produces attenuation coefficients within appropriate ranges while maintaining the time efficiency of direct PET-based generation
Solution Approach 2:
The patent implements a feedback mechanism where the generated attenuation coefficient image is evaluated for value validity. When values fall outside appropriate ranges, the system adjusts the generation process using the intermediate image guidance, creating a closed-loop system that ensures reliability while maintaining efficiency
2Device complexity
If a machine learning model directly outputs attenuation coefficient images from PET data, then the process is simplified, but there is no mechanism to ensure values remain within appropriate ranges
Solution Approach 1:
The patent segments the attenuation coefficient image generation process into two distinct stages: first generating an intermediate image containing tissue area information, then using this intermediate image to guide the generation of the final attenuation coefficient image. This segmentation allows each stage to focus on specific aspects, maintaining simplicity while improving precision
Solution Approach 2:
The patent performs preliminary action by generating the intermediate image before generating the final attenuation coefficient image. This intermediate image prepares tissue area information in advance, which then guides the final generation process to ensure values remain within appropriate ranges without adding significant complexity
Data Source
AI summary
This attenuation coefficient image generation method includes a step of generating an input image (6), a step of generating an intermediate image (7) including an image relating to tissue areas based on the input image (6), and a step of generating an attenuation coefficient image (9) based the intermediate image (7) and known attenuation coefficients of tissue areas.