Medical Image Super-Resolution via Deep Learning Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Medical images are often degraded due to imaging environment and system influences, necessitating effective processing to improve image quality for accurate diagnosis.
Innovation Solution
An image processing method using deep learning-based segmentation and super-resolution neural networks to acquire, segment, and reconstruct medical images, reducing computational load and enhancing image quality by isolating regions of interest and performing normalization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If deep learning-based segmentation and super-resolution are applied to the entire original image, then image quality enhancement is achieved, but computational load and processing time increase significantly
Solution Approach 1:
The patent divides the original image into multiple local regions and processes each region independently through deep learning-based segmentation and super-resolution. This segmentation approach reduces the computational complexity compared to processing the entire image at once, while still achieving high-quality enhancement in critical areas.
Solution Approach 2:
The patent applies different processing strategies to different regions of the image. By identifying and prioritizing certain regions for enhanced processing while reducing or skipping processing in less critical areas, the system achieves high image quality where needed while maintaining overall processing efficiency.
2Loss of information
If deep learning-based segmentation is performed on the entire image, then regions of interest are identified, but processing time and computational resources increase
Solution Approach 1:
The patent performs deep learning-based segmentation on divided local regions rather than the entire image at once. This approach maintains accurate region identification by focusing computational resources on smaller, manageable segments while reducing overall processing time through parallel processing of multiple regions.
Solution Approach 2:
The patent applies segmentation and processing to only certain regions of the image that are most relevant or critical, rather than uniformly processing the entire image. This partial action approach reduces processing time while maintaining sufficient information accuracy for diagnostic purposes.
Data Source
AI summary
An image processing device and an image processing method are disclosed. The image processing method comprises: acquiring an image; segmenting the acquired image by using a deep learning-based segmentation process to obtain a binarized image labeled with a region of interest; processing a pixel matrix of the acquired image by using a pixel matrix of the binarized image to obtain a segmented image; and performing super-resolution reconstruction on the segmented image by using a deep learning-based super-resolution neural network to obtain a reconstructed image.