Method for automatic contour extraction of a cardiac image
AssigneeUniversity of Cincinnati
InventorsChia Y. Han, David T. Porembka, Kwun-Nan Lin
Life of this patent
Estimating the term…
Abstract
An improved method is provided for extracting the edge boundary of an object presented as an ultrasound video image, in particular the edge boundary of the LV of the heart, which includes preprocessing algorithms, contour extraction algorithms and knowledge based contour following algorithms.
Claims26
A method for determining the boundary location of an object presented as a video image, the video image comprising a plurality of picture elements known as pixels, each pixel having a respective intensity level, the method comprising the steps of:
- (a) generating an original video image of the object;
- (b) creating a reduced size feature picture of the original video image of the object;
- (c) selecting a threshold value of pixel intensity;
- (d) setting the intensity level of all pixels of the feature picture which have a respective intensity level greater than said threshold value equal to a first predetermined intensity level;
- (e) setting the intensity level of all pixels of the feature picture which have a respective intensity level less than said threshold value equal to a second predetermined intensity level;
- (f) automatically identifying a plurality of adjacent pixels of the feature picture as corresponding to an internal region of the object;
- (g) automatically identifying a plurality of internal region boundary pixels as corresponding to the boundary of the internal region;
- (h) automatically identifying respective pixels of the original video image which correspond to the internal region boundary pixels identified in step (g);
- and
- (i) automatically identifying in the original video image respective object boundary pixels by selecting each respective object boundary pixel from a respective group of candidate pixels which contains a respective pixel identified in step (h).
A method of determining a threshold value for processing a video image, the video image comprising a plurality of picture elements known as pixels, each pixel having a respective intensity level, the method comprising the steps of:
- (a) computing a first histogram based on the pixels of the video image, said first histogram having an intensity level axis and a quantity of pixels axis;
- (b) identifying a plurality of pixels which comprise a predetermined percentage of the pixels of the video image having the highest intensity levels;
- (c) eliminating all pixels identified in step (b) from consideration;
- (d) computing a second histogram based solely on the remaining pixels of the video image which have not been eliminated from consideration, said second histogram having an intensity level axis and a quantity of pixels axis;
- (e) determining the highest number of pixels of any given intensity level;
- (f) determining the point on the descending side of the curve of the second histogram which corresponds to a second predetermined percentage of the number of pixels determined in step (e);
- (g) determining the point on the descending side of the curve of the second histogram which corresponds to a third predetermined percentage of the number of pixels determined in step (e);
- (h) determining the point on the intensity level axis of the second histogram which is collinear with the point determined in step (f) and the point determined in step (g);
- and
- (i) setting the threshold level equal to the intensity level corresponding to the intensity level of the point determined in step (h).
Dependencies are inferred from the claim text.