Method and system for machine learning classification based on structure or material segmentation in an image
Inventors
Assignees
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-12217418-B2
Publication Date
2025-02-04
Expiration Date
Abstract
A system and method for classifying a structure or material in an image of a subject. The system comprises: a segmenter configured to segment an image into one or more segmentations that correspond to respective structures or materials in the image, and to generate from the segmentations one or more segmentation maps of the image (each of the segmentation maps representing the image) including categorizations of pixels or voxels of the segmentation maps assigned from one or more respective predefined sets of categories; a classifier that implements a classification machine learning model configured to generate, based on the segmentations maps, one or more classifications and to assign to the classifications respective scores indicative of a likelihood that the structure or material, or the subject, falls into the respective classifications; and an output for outputting a result indicative of the classifications and scores.
Core Innovation
The invention provides a system for classifying a structure or material in an image of a subject. A segmenter segments the image into one or more segmentations corresponding to respective structures or materials and generates one or more segmentation maps, including categorizations of pixels or voxels assigned from respective predefined sets of categories.
The system further includes a classifier that implements a trained classification machine learning model. Based on the segmentation maps, the classifier generates one or more classifications and assigns respective scores indicative of a likelihood that the structure or material, or the subject, falls into the respective classifications, and an output outputs a result indicative of the classifications and scores.
In the segmenter, the structure segmenter generates structure segmentation maps using a structure segmentation machine learning model and categorizations from a predefined set of structure categories. The material segmenter generates material segmentation maps using a material segmentation machine learning model and categorizations from a predefined set of material categories.
The segmenter can also include an abnormality segmenter that generates abnormality segmentation maps using an abnormality segmentation model and categorizations from a predefined set of abnormality or normality categories.
Claims Coverage
The independent claims are clm-00001 and clm-00007. Across these, the claim set centers on a segmentation-first pipeline that produces segmentation maps with predefined category categorizations, followed by a trained classification machine learning model that outputs classifications with scores indicating likelihood.
Segmentation maps from predefined structure/material categories
A segmenter segments an image into one or more segmentations corresponding to respective structures or materials and generates one or more segmentation maps including categorizations of pixels or voxels assigned from one or more respective predefined sets of categories.
Trained classification model generating classifications with likelihood scores
A classifier implements a trained classification machine learning model configured to generate, based on the segmentation maps, one or more classifications and to assign to the classifications respective scores indicative of a likelihood that the structure or material, or the subject, falls into the respective classifications.
Outputting classifications and scores
An output outputs a result indicative of the classifications and scores.
Structure, material, and optional abnormality segmentation models
The segmenter comprises a structure segmenter configured to generate structure segmentation maps using a structure segmentation machine learning model and categorizations from a predefined set of structure categories, and a material segmenter configured to generate material segmentation maps using a material segmentation machine learning model and categorizations from a predefined set of material categories, and/or an abnormality segmenter configured to generate abnormality segmentation maps using an abnormality segmentation model and categorizations from a predefined set of abnormality or normality categories.
The claim coverage is built around generating segmentation maps with predefined category-based pixel or voxel categorizations, using a trained classification machine learning model to generate classifications and likelihood-indicative scores from the segmentation maps, and outputting a result containing the classifications and scores, with segmentation provided by structure, material, and optionally abnormality or normality models.
Stated Advantages
Documented Applications
Medical imaging-based outputs including fracture risk probabilities over multiple timeframes.
Alternative medical application outputs including disease progression states, treatment efficacy, and condition likelihoods.
Medical condition-related outputs for osteomalacia, tumour, osteonecrosis, and infection.
Interested in licensing this patent?