Method and apparatus for generating 3D knee joint image

Patent metadata

Field	Value
Publication number	US-9349074-B2
Application number	US-201314142388-A
Country	US
Kind code	B2
Filing date	Dec 27, 2013
Priority date	Oct 25, 2013
Publication date	May 24, 2016
Grant date	May 24, 2016

Abstract

Disclosed herein are a method and an apparatus for generating a 3D knee joint image. According to an exemplary embodiment of the present invention, a method for generating a 3D knee joint image includes: generating a statistical shape model by using previously generated sample knee bone images; generating a knee joint bone image by segmenting a bone region from an input knee image using the statistical shape model; generating a keen joint cartilage image by segmenting a cartilage region from the knee joint bone image; and generating a 3D knee joint image for the input knee image by composing the knee joint bone image with the knee joint cartilage image. According to the exemplary embodiments of the present invention, it is possible to more rapidly and accurately generate the 3D knee joint image.

First claim

What is claimed is:

1. A method for generating a 3D knee joint image, comprising:
generating a statistical shape model by using previously generated sample knee bone images;
generating a knee joint bone image by segmenting a bone region from an input knee image using the statistical shape model;
generating a knee joint cartilage image by segmenting a cartilage region from the knee joint bone image; and
generating a 3D knee joint image for the input knee image by composing the knee joint bone image with the knee joint cartilage image,
wherein the generating of the statistical shape model includes:
generating matching images for each of the sample knee bone images by using original images and reduction images of each of the sample knee bone images;
generating final matching images by using the matching images;
generating feature point vectors for each of the sample bone knee images by extracting feature points of the final matching images; and
generating the statistical shape model by using an average vector and a covariance matrix of the feature point vectors.

2. The method of claim 1 , wherein the generating of the matching images includes:
extracting feature points of the original images and feature points of the reduction images;
extracting shape contexts for the feature points of the original images and shape contexts for the feature points of the reduction images; and
generating the matching images by matching the reduction images with the original images using the shape contexts of the original images and the shape contexts of the reduction images.

3. The method of claim 1 , wherein the generating of the final matching images includes:
determining a reference matching image and a non-reference matching image among the matching images;
extracting feature points of the reference matching image and feature points of the non-reference image;
extracting shape contexts for the feature points of the reference matching image and shape contexts for the feature points of the non-reference matching image; and
generating the final matching images by matching the non-reference matching image with the reference matching image using the shape contexts of the reference matching image and the shape contexts of the non-reference matching image.

4. A method for generating a 3D knee joint image, comprising:
generating a statistical shape model by using previously generated sample knee bone images;
generating a knee joint bone image by segmenting a bone region from an input knee image using the statistical shape model;
generating a knee joint cartilage image by segmenting a cartilage region from the knee joint bone image; and
generating a 3D knee joint image for the input knee image by composing the knee joint bone image with the knee joint cartilage image,
wherein the generating of the knee joint bone image includes:
disposing the statistical shape model on the input knee image;
calculating a brightness value vector of a point, at which each of the feature points of the statistical shape model is positioned, in the input knee image;
moving each of the feature points by using the brightness value vector; and
segmenting a bone region from the input knee image by using each of the moving feature points.

5. A method for generating a 3D knee joint image, comprising:
generating a statistical shape model by using previously generated sample knee bone images;
generating a knee joint bone image by segmenting a bone region from an input knee image using the statistical shape model;
generating a knee joint cartilage image by segmenting a cartilage region from the knee joint bone image; and
generating a 3D knee joint image for the input knee image by composing the knee joint bone image with the knee joint cartilage image,
wherein the generating of the knee joint cartilage image includes:
setting an interest region in the knee joint bone image;
extracting a bone-cartilage interface from the interest region; and
segmenting a cartilage region from the knee joint bone image by using brightness value information of the bone-cartilage interface.

Assignees

• Kwangju Inst Sci & Tech

Inventors

• Choi Tae Sun
• Choi Wook-Jin

Classifications

• G06F18/22

  Matching criteria, e.g. proximity measures · CPC title

• G06T7/12 Primary

  Edge-based segmentation · CPC title

• G06T2207/10088

  Magnetic resonance imaging [MRI] · CPC title

• G06T7/0083

  Physics · mapped topic

• G06T7/0089

  Physics · mapped topic