Masking Unwanted Pixels in Medical Images via Texture Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing techniques are ineffective in reliably removing unwanted pixels from medical images, particularly in phase images where intensity-based thresholding fails due to extreme phase values and lack of correlation with neighboring pixels.
Innovation Solution
The method involves texture analysis using grey level co-occurrence matrix statistics to identify highly textured regions in images, generating a mask to exclude these regions, and applying it to the image to create a masked version that focuses on regions of interest.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If intensity-based thresholding is used to remove unwanted pixels, then it works effectively for magnitude images, but it fails for phase images due to extreme phase values and lack of correlation with neighboring pixels
Solution Approach 1:
The patent changes the parameter used for pixel selection from intensity-based thresholding to texture-based analysis. By computing texture features (such as local gradient variations, entropy, or co-occurrence matrix statistics) for each pixel and comparing them against thresholds, the method adapts to both magnitude and phase images, resolving the contradiction between reliability for magnitude images and adaptability to different image types.
2Measurement precision
If unwanted pixels are removed from phase images, then image processing accuracy improves, but computational complexity increases due to the need for texture analysis
Solution Approach 1:
The patent segments the image processing task into distinct steps: computing texture features for each pixel, identifying unwanted pixels based on texture thresholds, generating a mask, and applying the mask. This segmentation allows for optimized computation at each stage and improves measurement precision by systematically eliminating unwanted pixels while managing computational complexity through structured processing.
Solution Approach 2:
The patent performs preliminary texture analysis and feature computation before the main image processing task. By pre-computing texture features and identifying unwanted pixels in advance, the method reduces the computational burden during subsequent analysis steps and improves overall processing efficiency while maintaining high accuracy.
3Reliability
If texture analysis is performed to identify unwanted pixels, then removal effectiveness improves, but processing time increases
Solution Approach 1:
The patent applies partial texture analysis by computing texture features only for pixels that meet certain preliminary criteria or by using simplified texture metrics for initial screening. This partial action approach maintains reliable unwanted pixel removal effectiveness while reducing the overall processing time compared to performing comprehensive texture analysis on every pixel without filtering.
Data Source
AI summary
According to an aspect, there is provided an apparatus (300) for masking textured regions in an image. The apparatus (300) comprises a processor (302) configured to receive an input image having a plurality of pixels; determine, for each pixel of the plurality of pixels, a first image texture feature value; generate, based on the first image texture feature values, a mask; and apply the mask to the received input image to generate a masked image. A computer-implemented method and a computer program product are also provided.