Method and system of classifying medical images

• US 9,122,955 B2
• Filed: 06/28/2011
• Issued: 09/01/2015
• Est. Priority Date: 06/28/2010
• Status: Expired due to Fees

• Current Assignee: BAR ILAN University, Ramot at Tel Aviv University Limited (Tel Aviv University), Tel Hashomer Medical Research Infrastructure And Services Ltd.
• Original Assignee: BAR ILAN University, Ramot at Tel Aviv University Limited (Tel Aviv University), Tel Hashomer Medical Research Infrastructure And Services Ltd.
• Inventors: Goldberger, Jacob, Greenspan, Hayit, Avni, Uri, Konen, Eli, Sharon, Michal
• Primary Examiner(s): Koziol, Stephen R
• Assistant Examiner(s): CONNER, SEAN M
• Application Number:
  US13/170,200
• Publication Number:
  US 20110317892A1
• Time in Patent Office: 1,526 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: A61B 6/503 for diagnosis of the heart
• CPC Class Codes: A61B 6/5217 extracting a diagnostic or physiological parameter from medical diagnostic data
• CPC Class Codes: G06F 16/51 Indexing; Data structures therefor; Storage structures
• CPC Class Codes: G06F 16/5866 using information manually generated, e.g. tags, keywords, comments, manually generated location and time information
• CPC Class Codes: G06F 18/214 Generating training patterns; Bootstrap methods, e.g. bagging or boosting
• CPC Class Codes: G06F 18/23 Clustering techniques
• CPC Class Codes: G06F 18/24 Classification techniques
• CPC Class Codes: G06T 2207/30048 Heart; Cardiac
• CPC Class Codes: G06T 2207/30061 Lung
• CPC Class Codes: G06T 3/10 Selection of transformation methods according to the characteristics of the input images
• CPC Class Codes: G06T 7/0012 Biomedical image inspection
• CPC Class Codes: G06V 10/464 using a plurality of salient features, e.g. bag-of-words [BoW] representations

Abstract

A method of generating a category model for classifying medical images. The method comprises providing a plurality of medical images each categorized as one of a plurality of categorized groups, generating an index of a plurality of visual words according to a distribution of a plurality of local descriptors in each the image, modeling a category model mapping a relation between each visual word and at least one of the categorized groups according to the index, and outputting the category model for facilitating the categorization of an image based on local descriptors thereof.

Claims

1. A method of generating a category model for classifying medical images, comprising:

providing a plurality of anatomical medical images imaging at least one internal organ or area of a body;

in each of said plurality of anatomical medical images identifying a plurality of image patches, each image patch is represented by a plurality of repeatable multidimensional features in a pixel area of a respective anatomical medical image and image coordinates of said pixel area in said respective anatomical medical image, wherein said plurality of repeatable multidimensional features of each of the plurality of image patches comprises at least one noise reduction based coefficient and a gray level feature of the image patch;

for each of said image patches, weighting at least some of the noise reduction based coefficient, the gray level feature, and the image coordinates, wherein said weighting is tuned experimentally on a cross-validation set;

generating a plurality of visual words by clustering said plurality of image patches represented by said weighted at least one noise reduction based coefficient, gray level feature, and image coordinates of each one of said plurality of image patches;

modeling a category model mapping a relation between said plurality of visual words;

outputting said category model adapted to categorize a pathology in a new anatomical medical image based on a new set of image patches from said new anatomical medical image and the coordinates of pixel areas of said new set of image patches in said new anatomical medical image,

wherein at least one of said providing, generating, modeling, and outputting is performed by at least one processor.

2. The method of claim 1, further comprising dividing said plurality of medical images among a plurality of categorized groups.

3. The method of claim 1, wherein an index of said plurality of visual words comprises less than 700 visual words.

4. The method of claim 1, wherein said plurality of medical images are part of a training set having more than 10,000 medical images.

5. The method of claim 1, wherein said generating comprises clustering a plurality of local descriptors based on repeatable multidimensional features in each image in a plurality of clusters, said plurality of visual words being defined according to said plurality of clusters.

6. The method of claim 5, wherein said clustering is performed according to a principal component analysis (PCA).

7. The method of claim 1, wherein said modeling is performed using a support vector machine (SVM) training procedure.

8. The method of claim 7, wherein said SVM training procedure is a multi-class SVM with a radial basis function (RBF) kernel.

9. The method of claim 1, wherein said plurality of medical images are provided from a picture archiving communication system (PACS).

10. The method of claim 1, wherein a plurality of categorized groups define a plurality of said pathologies.

11. The method of claim 1, further comprising automatically categorizing said plurality of medical images.

12. The method of claim 1, wherein the category model is updated upon each usage of the category model.

13. The method of claim 1, further comprising normalizing each image patch, wherein each normalized image patch is formed from a transformation of intensity values from a corresponding image patch, to render the image patch less variant to brightness, and to provide local contrast enhancement.

14. The method of claim 13, wherein said intensity values from the image patch are obtained from pixels of the image patch.

15. The method of claim 1, wherein said repeatable multidimensional features in each said image are from three dimensional images.

16. The method of claim 1, wherein said outputting comprises outputting said category model for facilitating the categorization of an image based on local descriptors thereof including said image from three dimensional images.

17. The method of claim 10, wherein the plurality of pathologies is selected from the group consisting of enlarged heart, lung infiltrates, right pleural effusion and left pleural effusion.