Method for automated delineation of contours of tissue in medical images
AssigneeThe University of British Columbia University-Industry Liaison Office
InventorsSara Badiei, Septimiu Edmund Salcudean, Seyedeh Sara Mahdavi
Life of this patent
Estimating the term…
Abstract
The present invention relates to a method for delineating the contour of an object in captured medical images by first transforming the shape of the object into a simple geometric shape that is more computationally tractable than the shape of the object. After the contour of the transformed shape is detected, the inverse of the transformation is applied to the contour such that it represents the contour of the object in the captured medical image.
Claims20
A computer readable medium having instructions and statements stored therein for execution by a processor to perform a method for delineating the contour of a tissue region in a captured medical image, the captured image comprising a captured cross-section of the tissue region, the method comprising the steps of:
- (a) generating a transformed image of the captured image by applying a transformation to the captured image such that the captured cross-section in the captured image is transformed into a transformed cross-section in the transformed image that resembles a pre-determined two-dimensional shape which is more computationally tractable than the captured cross-section;
- (b) solving for an initial edge contour of the tissue region by fitting the pre-determined two-dimensional shape to the transformed cross-section;
- (c) detecting the edge of the transformed cross-section by applying an edge-detector method within a search area of the transformed image about the initial edge contour;
- (d) solving for a final edge contour by fitting the pre-determined two-dimensional shape to the detected edge of the transformed cross-section;
- (e) solving for the two-dimensional contour of the tissue region in the captured image by applying the inverse of the transformation to the final edge contour;
- and
- (f) displaying the two-dimensional contour on a display.
A computer readable medium having statements and instructions stored therein for execution by a processor to perform a method for delineating the contour of a tissue region in a set of captured medical images, each captured image comprising a captured cross-section of the tissue region, the method comprising the steps of:
- (a) selecting a first captured image from the set of captured images of the tissue region;
- (b) for the first captured image:
- (i) generating a transformed image of the first captured image by applying a transformation to the first captured image such that the captured cross-section in the first captured image is transformed into a transformed cross-section in the transformed image that resembles a pre-determined two-dimensional shape which is more computationally tractable than the captured cross-section;
- (ii) solving for an initial edge contour by fitting the pre-determined two-dimensional shape to the transformed cross-section;
- (iii) detecting the edge of the transformed cross-section by applying an edge-detector method within a search area of the transformed image about the initial edge contour;
- and
- (iv) solving for a final edge contour by fitting the pre-determined two-dimensional shape to the detected edge of the transformed cross-section;
- (c) for each captured image in the set of captured images, other than the first captured image:
- (i) generating a transformed image by applying the transformation to the captured image such that the captured cross-section in the captured image is transformed into a transformed cross-section in the transformed image that resembles the pre-determined two-dimensional shape;
- (ii) solving for an initial edge contour;
- (iii) detecting the edge of the transformed cross-section by applying an edge-detector method within a search area of the transformed image about the initial edge contour;
- and
- (iv) solving for the final edge contour by fitting the first pre-determined two-dimensional shape to the detected edge of the transformed cross-section;
- (d) solving for a final surface contour of the tissue region by fitting a pre-determined three-dimensional shape to the final edge contours associated with each captured image in the set of captured images;
- (e) solving a three-dimensional contour of the tissue region by applying the inverse of the transformation to the final surface contour;
- and
- (f) displaying a cross-section of the three-dimensional contour on a display.
Dependencies are inferred from the claim text.