Back to results
US 2012/0237107 A1Published application
System and methods for digital evaluation of cellblock preparations
AssigneeUniversity of Kansas
InventorsOssama Tawfik, Brian Potetz
Life of this patent
Estimating the term…
Abstract
Digital evaluation of cellblock preparations to determine the type and extent of disease in order to identify the best approach for treatment without the need for additional testing or sampling.
Claims20
A computer system method for evaluation of cellblock preparations, comprising the steps of:
- producing digitally a treated Pap smear sample to obtain a digital image;
- segmenting the digital image to obtain a plurality of segmented digital images;
- processing each segmented digital image of the plurality, wherein said processing step further comprises the steps of:
- locating a nucleus of a cell within each segmented digital image, extracting one or more visual features of the nucleus, and ranking the one or more visual features of the nucleus to obtain one or more quantified features within each segmented digital image;
- joining all segmented digital images to obtain a final digital image;
- collecting all quantified features of the final digital image to obtain a collection of quantified features;
- correlating the collection of quantified features to a diagnosis classification;
- and
- communicating the diagnosis classification.
6 more elements
A system for evaluation of cellblock preparations, comprising:
- a feature extraction component, wherein said feature extraction component receives a digital Pap smear image, segments the digital Pap smear image to obtain a plurality of segmented images, locates a nucleus of a cell within each segmented image and ranks one or more visual features of the nucleus to obtain one or more quantified features within each segmented image;
- and
- a classification component, wherein said classification component joins all segmented images to obtain a final digital image, collects the one or more quantified features of the final digital image to obtain a collection of quantified features, and correlates the collection of quantified features to a diagnosis classification in order to communicate the diagnosis classification.
Dependencies are inferred from the claim text.