Medical Image Segmentation Masks Using Parallel Region Classifiers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing segmentation algorithms for medical imaging are slow and provide limited resolution, hindering their utility in applications such as quantitative analysis and medical abnormality identification.
Innovation Solution
A framework that generates segmentation mask data using a trained machine learning model, where a first descriptor is obtained for a location in medical imaging data, and class labels are determined for regions using multiple classifiers, allowing for fast and high-resolution segmentation mask generation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If existing segmentation algorithms are used, then segmentation masks can be generated, but the generation speed is slow and resolution is limited
Solution Approach 1:
The patent divides the image processing task into multiple parallel classifiers, each handling specific regions. The image is processed by obtaining descriptors for multiple locations simultaneously, with each location having multiple classifiers that process different regions in parallel, thereby increasing generation speed while maintaining high resolution through the coordinated output of all classifiers
Solution Approach 2:
The patent transitions from traditional single-pass segmentation to a multi-dimensional approach by creating a classifier network with multiple locations and multiple regions per location. This dimensional expansion allows simultaneous processing of multiple image regions through parallel classifier execution, resolving the speed-resolution tradeoff
2Loss of time
If existing segmentation algorithms are used, then segmentation masks can be generated, but the processing time is excessive
Solution Approach 1:
The patent performs preliminary actions by pre-defining multiple locations and their associated regions and classifiers before processing the actual image. The framework pre-establishes the classifier network structure with all locations and regions configured in advance, enabling immediate parallel processing when image data is input, thereby reducing processing time while maintaining high productivity
Solution Approach 2:
The patent ensures continuous useful action by having all classifiers process their respective regions simultaneously and continuously throughout the segmentation process. Rather than sequential processing, the system maintains continuous parallel computation across all locations and regions, eliminating idle time and maximizing both speed and productivity
Data Source
AI summary
A framework for generating segmentation mask data for first medical imaging data. The framework may include obtaining a first descriptor for a first location in the first medical imaging data, the first descriptor being representative of values of elements of the first medical imaging data located relative to the first location according to a first predefined pattern. Based on an input of the first descriptor to a trained machine learning model, a class label may be determined for each of a plurality of regions of the first medical imaging data, each region having a respective different predetermined location relative to the first location, the class label for each one of the plurality of regions being determined using a respective different one of a plurality of classifiers of the trained machine learning model. The segmentation mask data may be generated for the first medical imaging data based on the class labels determined for the plurality of regions.