Segmentation of regions in measurements of a body based on a deformable model

A method for segmenting significant regions in measurements of a target body, comprising the steps of:

• receiving reference segmentation data comprising
• a plurality of coordinate values for spatial or temporal dimensions for each vertex in a first plurality of vertices that indicates a first boundary of a myocardial wall of a body, and
• a second plurality of coordinate values for spatial or temporal dimensions for each vertex in a second plurality of vertices that indicates a second boundary of the myocardial wall of the body;
• receiving target data comprising a plurality of values each associated with a plurality of coordinate values for spatial or temporal dimensions based on a measurement mode of the body;
• and
• adjusting a coordinate for a first vertex of the first plurality of vertices toward a higher gradient in values of the target data in a vicinity of the first vertex, comprising the steps of
• determining a local separation distance from the first vertex to the second boundary,
• adjusting the coordinate based on the local separation distance.

A method for segmenting significant regions in measurements of a target body, comprising the steps of:

• receiving reference segmentation data comprising a plurality of coordinate values for spatial or temporal dimensions for each vertex in a plurality of vertices that indicates a boundary of a physical component of a body;
• receiving target data comprising a plurality of values each associated with a plurality of coordinate values for spatial or temporal dimensions based on a measurement mode of the body;
• and
• adjusting a coordinate for a vertex of the plurality of vertices toward a higher gradient in values of the target data in a vicinity of the vertex, further comprising the steps of
• computing a measure of curvature at the vertex,
• determining whether the measure of curvature exceeds a threshold curvature, and
• if it is determined that the measure of curvature exceeds the threshold curvature, then changing a coordinate of the vertex to reduce the measure of curvature,
• whereby after all adjustments in an absence of a higher gradient, the plurality of vertices form a non-circular shape in at least one plane.

A method for segmenting significant regions in measurements of a target body, comprising the steps of:

• receiving segmentation data comprising a plurality of coordinate values for spatial or temporal dimensions for each vertex in a plurality of vertices that indicates a boundary of a physical component of a body;
• receiving target data comprising a plurality of values each associated with a plurality of coordinate values for spatial or temporal dimensions based on a measurement mode of the body;
• and
• adjusting a coordinate for a vertex of the plurality of vertices toward a higher gradient in values of the target data in a vicinity of the vertex, further comprising
• deriving a vector field that points toward high gradient coordinates in the target data which represent edges in the target data based on applying in three dimensions a generalized gradient vector field (GGVF) process that iteratively grows from an initial vector field based on edge detection in the target data to the vector field by filling regions of low amplitude edges with vectors that point to a region of large amplitude edges;
• and
• adjusting the coordinate based on the vector field.

A method for segmenting significant regions in measurements of a target body, comprising the steps of:

• receiving reference range data that indicates a distribution of distances between adjacent vertices among all pairs of adjacent vertices in a reference plurality of reference vertices that indicates a reference boundary of a physical component in a reference body;
• receiving segmentation data comprising a plurality of coordinate values for spatial or temporal dimensions for each vertex in a particular plurality of vertices that indicates a particular boundary of the physical component of a particular body;
• receiving target data comprising a plurality of values each associated with a plurality of coordinate values for spatial or temporal dimensions based on a measurement mode of the particular body;
• and
• adjusting a coordinate for a vertex of the particular plurality of vertices toward a higher gradient in values of the target data in a vicinity of the vertex, comprising the steps of
• determining a distance from the vertex to an adjacent vertex of the first plurality of vertices,
• determining whether the distance is within a range of distances based on the reference range data, and
• if it is determined that the distance is not within the range of distances, then determining an edge length adjustment proportional to a difference between the distance and the range of distances.

A non-transitory computer-readable medium carrying one or more sequences of instructions for segmenting significant regions in measurements of a target body, wherein execution of the one or more sequences of instructions by one or more processors causes the one or more processors to perform the steps of:

• receiving reference data comprising a plurality of values each associated with a plurality of coordinate values for spatial or temporal dimensions based on a first measurement mode of a reference body;
• receiving reference segmentation data comprising a plurality of coordinate values for spatial or temporal dimensions for each vertex in a plurality of vertices that indicates a boundary of a physical component of the reference body in the same coordinate frame as the reference measured data;
• receiving target data comprising a plurality of values each associated with a plurality of coordinate values for spatial or temporal dimensions based on a second measurement mode of a target body;
• automatically determining a particular transformation to maximize a similarity measure between the reference data and the target data;
• transforming the reference segmentation data using the particular transformation to produce transformed segmentation data that indicates in the target data a boundary of the physical component of the target body, and
• adjusting a coordinate for a vertex of the transformed segmentation data based on a plurality of gradients in values of the target data in a vicinity of the vertex.

A non-transitory computer-readable medium carrying one or more sequences of instructions for segmenting significant regions in measurements of a target body, wherein execution of the one or more sequences of instructions by one or more processors causes the one or more processors to perform the steps of:

• receiving reference segmentation data comprising
• a plurality of coordinate values for spatial or temporal dimensions for each vertex in a first plurality of vertices that indicates a first boundary of a myocardial wall of a body, and
• a second plurality of coordinate values for spatial or temporal dimensions for each vertex in a second plurality of vertices that indicates a second boundary of the myocardial wall of the body;
• receiving target data comprising a plurality of values each associated with a plurality of coordinate values for spatial or temporal dimensions based on a measurement mode of the body;
• and
• adjusting a coordinate for a first vertex of the first plurality of vertices toward a higher gradient in values of the target data in a vicinity of the first vertex, comprising the steps of
• determining a local separation distance from the first vertex to the second boundary,
• adjusting the coordinate based on the local separation distance.

A non-transitory computer-readable medium carrying one or more sequences of instructions for segmenting significant regions in measurements of a target body, wherein execution of the one or more sequences of instructions by one or more processors causes the one or more processors to perform the steps of:

• receiving reference segmentation data comprising a plurality of coordinate values for spatial or temporal dimensions for each vertex in a plurality of vertices that indicates a boundary of a physical component of a body;
• receiving target data comprising a plurality of values each associated with a plurality of coordinate values for spatial or temporal dimensions based on a measurement mode of the body;
• and
• adjusting a coordinate for a vertex of the plurality of vertices toward a higher gradient in values of the target data in a vicinity of the vertex, further comprising the steps of
• computing a measure of curvature at the vertex,
• determining whether the measure of curvature exceeds a threshold curvature, and
• if it is determined that the measure of curvature exceeds the threshold curvature, then changing a coordinate of the vertex to reduce the measure of curvature,
• whereby after all adjustments in an absence of a higher gradient, the plurality of vertices form a non-circular shape in at least one plane.

A non-transitory computer-readable medium carrying one or more sequences of instructions for segmenting significant regions in measurements of a target body, wherein execution of the one or more sequences of instructions by one or more processors causes the one or more processors to perform the steps of:

• receiving segmentation data comprising a plurality of coordinate values for spatial or temporal dimensions for each vertex in a plurality of vertices that indicates a boundary of a physical component of a body;
• receiving target data comprising a plurality of values each associated with a plurality of coordinate values for spatial or temporal dimensions based on a measurement mode of the body;
• and
• adjusting a coordinate for a vertex of the plurality of vertices toward a higher gradient in values of the target data in a vicinity of the vertex, further comprising
• deriving a vector field that points toward high gradient coordinates in the target data which represent edges in the target data based on applying in three dimensions a generalized gradient vector field (GGVF) process that iteratively grows from an initial vector field based on edge detection in the target data to the vector field by filling regions of low amplitude edges with vectors that point to a region of large amplitude edges;
• and
• adjusting the coordinate based on the vector field.

A non-transitory computer-readable medium carrying one or more sequences of instructions for segmenting significant regions in measurements of a target body, wherein execution of the one or more sequences of instructions by one or more processors causes the one or more processors to perform the steps of:

• receiving reference range data that indicates a distribution of distances between adjacent vertices among all pairs of adjacent vertices in a reference plurality of reference vertices that indicates a reference boundary of a physical component in a reference body;
• receiving segmentation data comprising a plurality of coordinate values for spatial or temporal dimensions for each vertex in a particular plurality of vertices that indicates a particular boundary of the physical component of a particular body;
• receiving target data comprising a plurality of values each associated with a plurality of coordinate values for spatial or temporal dimensions based on a measurement mode of the particular body;
• and
• adjusting a coordinate for a vertex of the particular plurality of vertices toward a higher gradient in values of the target data in a vicinity of the vertex, comprising the steps of
• determining a distance from the vertex to an adjacent vertex of the first plurality of vertices,
• determining whether the distance is within a range of distances based on the reference range data, and
• if it is determined that the distance is not within the range of distances, then determining an edge length adjustment proportional to a difference between the distance and the range of distances.