Curvature based system for the segmentation and analysis of image data

A method for segmentation of a discrete pixel image, the image being defined by data representative of characteristics of a plurality of pixels of an image matrix, the method comprising the steps of:

• computing gradients for an image characteristic;
• storing the gradients;
• initializing a candidate boundary;
• and
• evolving the candidate boundary to a boundary of a feature of interest based upon the gradients and mean curvature weighted normals for the candidate boundary.

A method for analyzing a heart from a discrete pixel image, the method comprising the steps of:

• computing gradients for a characteristic of pixels defining the image;
• storing the gradients;
• initializing a candidate boundary for measurement of at least the left ventricle of the heart;
• evolving the candidate boundary to a boundary at least the left ventricle by reference to the gradients and to mean curvature weighted normals;
• computing a volume of at least a portion of the left ventricle based the evolved candidate boundary.

A method for determining measuring a cardiac parameter from a series of images acquired at different locations along a left ventricle of the heart and at different times during a cardiac cycle, the method comprising the steps of:

• analyzing image data for the series of images to determine respective gradient values of intensity values for at least a portion of the images including a feature of a heart;
• storing the gradient values;
• evolving a candidate boundary in each image to identify boundaries of a feature of interest based upon the gradient values and upon mean curvature weighted normals for the candidate boundary;
• and
• computing a measurement based upon the candidate boundaries.

An imaging and segmentation system comprising:

• an image data acquisition unit, the unit generating image data of a subject including a feature of interest;
• a control circuit coupled to the image data acquisition unit and configured to command acquisition of the image data;
• a memory circuit for storing the image data;
• and
• a segmentation and analysis circuit coupled to the memory circuit for identifying the feature of interest by generating gradients of a characteristic of the image data and for initiating and evolving a candidate boundary to a boundary of a feature of interest based upon the gradients and upon mean curvature weighted normals for the candidate boundary.