Back to results
US 9,286,672 B2Granted
Integrated multivariate image-based method for disease outcome predicition
AssigneeRutgers, The State University of New Jersey
InventorsAnant Madabhushi, Ajay Basavanhally
Life of this patent
Estimating the term…
Abstract
The described invention provides a system and method for predicting disease outcome using a multi-field-of-view scheme based on image-based features from multi-parametric heterogenous images.
Claims20
A method for analyzing a heterogenous image, using a multi-field-of-view (FOV) multi-parametric scheme, the method comprising:
- (a) inputting, via a computing device, a large heterogenous image;
- (b) generating, via the computing device, a plurality of FOVs at a first fixed FOV size from among a plurality of fixed FOV sizes;
- (c) producing, via the computing device, an aggregated multi-FOV decision for the large heterogenous image based on a first image parameter from among a plurality of image parameters, comprising:
- (i) producing a combined class decision for each of the plurality of FOVs at the first fixed FOV size based on the first image parameter, comprising:
- (1) extracting, via the computing device, the first image parameter describing a spatial arrangement of at least one object from each FOV at the first fixed FOV size;
- (2) training, via the computing device, a first classifier from among a plurality of classifiers identifying the first image parameter describing the spatial arrangement of the at least one object from each FOV at the first fixed FOV size;
- (3) making, via the computing device, a first class decision for each FOV at the first fixed FOV size using a classification ensemble based on the first image parameter;
- (4) combining, via the computing device, the first class decisions for each of the plurality of FOVs at the first fixed FOV size to yield the combined class decision for the plurality of FOVs at the first fixed FOV size;
- and
- (5) producing, via the computing device, the combined class decision for each of the plurality of FOVs at the first fixed FOV size based on the first image parameter;
- (ii) repeating, via the computing device, the steps of extracting, training, making and combining for each of the plurality of fixed FOV sizes to generate a plurality of combined class decisions at respective ones of the plurality of fixed FOV sizes based on the first image parameter;
- (iii) aggregating, via the computing device, the plurality of combined class decisions at the respective ones of the plurality of fixed FOV sizes based on the first image parameter to yield an aggregated multi-FOV decision, using a first multi-FOV classifier from among a plurality of multi-FOV classifiers, wherein the first multi-FOV classifier aggregates the plurality of combined class decisions that are outputs of the ensembles at each of the plurality of fixed FOV sizes based on the first image parameter;
- and
- (iv) producing the aggregated multi-FOV decision for the large heterogenous image based on the first image parameter.
11 more elements
A system for analyzing a heterogenous image, using a multi-fields of view (FOVs) multi-parametric scheme via a computing device, comprising:
- (a) a computing device configured to input an image;
- (b) the computing device configured to generate a plurality of FOVs at a first fixed FOV size from among a plurality of fixed FOV sizes;
- (c) the computing device configured to produce an aggregated multi-FOV decision for the large heterogenous image based on a first image parameter from among a plurality of image parameters, comprising:
- (i) the computing device configured to produce a combined class decision for each of the plurality of FOVs at the first fixed FOV size based on the first image parameter, comprising:
- (1) the computing device configured to extract the first image parameter describing a spatial arrangement of at least one object from each FOV at the first fixed FOV size;
- (2) the computing device configured to train a first classifier from among a plurality of classifiers identifying the first image parameter describing the spatial arrangement of the at least one object from each FOV at the first fixed FOV size;
- (3) the computing device configured to make a first class decision for each FOV at the first fixed FOV size using a classification ensemble based on the first image parameter;
- (4) the computing device configured to combine the first class decisions for each of the plurality of FOVs at the first fixed FOV size to yield the combined class decision for the plurality of FOVs at the first fixed FOV size;
- and
- (5) the computing device configured to produce the combined class decision for each of the plurality of FOVs at the first fixed FOV size based on the first image parameter;
- (ii) the computing device configured to repeat the steps of extracting, training, making and combining for each of the plurality of fixed FOV sizes to generate a plurality of combined class decisions at respective ones of the plurality of fixed FOV sizes based on the first image parameter;
- (iii) the computing device configured to aggregate the plurality of combined class decisions at the respective ones of the plurality of fixed FOV sizes based on the first image parameter to yield an aggregated multi-FOV decision, using a first multi-FOV classifier from among a plurality of multi-FOV classifiers, wherein the first multi-FOV classifier aggregates the plurality of combined class decisions that are outputs of the ensembles at each of the plurality of fixed FOV sizes based on the first image parameter;
- and
- (iv) the computing device configured to produce the aggregated multi-FOV decision for the large heterogenous image based on the first image parameter.
11 more elements
Dependencies are inferred from the claim text.