MRI Background Pixel Suppression via Deep Learning Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Magnetic resonance imaging (MRI) systems face challenges in efficiently processing background noise, which occupies a significant portion of the image and does not contain clinically relevant information, leading to increased processing time and noise enhancement.
Innovation Solution
A method and system utilizing a deep learning module to identify and differentiate background pixels from foreground pixels in MRI images, allowing for the suppression of background pixels during the image reconstruction process, thereby processing only the clinically relevant foreground pixels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If background pixels are processed during image reconstruction, then complete image data is maintained, but processing time increases and noise is enhanced
Solution Approach 1:
The patent extracts and removes background pixels from the image data before processing. A mask image is generated identifying background regions, and these pixels are excluded from subsequent reconstruction and filtering operations, thereby reducing processing time and noise while maintaining anatomical integrity
Solution Approach 2:
The patent segments the image into foreground (anatomical) and background (non-anatomical) regions using a mask image. This segmentation allows selective processing of only relevant foreground pixels, avoiding unnecessary computation on background pixels and reducing overall processing time
2Reliability
If background pixels are processed during image reconstruction, then complete image data is maintained, but computational load increases
Solution Approach 1:
The patent extracts background pixels and removes them from the processing pipeline. By excluding background pixels from reconstruction and filtering operations, the computational load is significantly reduced while maintaining image quality through focused processing on anatomical regions
Solution Approach 2:
Instead of processing all pixels (excessive action), the patent applies partial action by processing only the foreground pixels that contain anatomical information. This selective processing reduces computational requirements while maintaining sufficient image quality for diagnostic purposes
3Object-affected harmful factors
If denoising techniques are applied to remove background noise, then noise is reduced, but anatomical regions may be affected
Solution Approach 1:
The patent segments the image into background and foreground regions using a mask image. By applying segmentation before denoising, the system can selectively apply noise reduction only to background pixels while preserving anatomical regions, thus removing background noise without affecting anatomical integrity
Solution Approach 2:
The patent applies local quality by treating background and foreground regions differently. Background pixels undergo noise reduction processing while foreground anatomical pixels are preserved, achieving localized noise removal that maintains anatomical integrity and avoids artifacts
Data Source
AI summary
Method (1000) and system (100) for image processing for a medical device is provided. The method (1000) includes acquiring (1010) a plurality of images of a subject using an image acquisition system (110) of the medical device. The method (1000) further includes identifying and differentiating (1020) a plurality of background pixels and a plurality of foreground pixels in the images using the deep learning module (125). The method (1000) further includes suppressing (1030) the identified background pixels using a mask and processing the foreground pixels for subsequent reconstruction and/or visualization tasks.