Method for local surface smoothing with application to chest wall nodule segmentation in lung CT data

A computer-readable medium having program instructions stored thereon for execution by a processor to perform method of local surface smoothing of a three-dimensional (“3D”) binary volume of interest (“VOI”), wherein the VOI comprises a plurality of 3D surface points, the method comprising:

• identifying a projection plane;
• obtaining a two-dimensional (“2D”) projection image by projecting the plurality of 3D surface points onto the projection plane;
• determining a plurality of intensities of the projection image, the plurality of intensities comprising distances from the plurality of 3D surface points to the projection plane;
• identifying a bump region on the 2D projection image, wherein the plurality of 3D surface points comprises points outside and around the bump region and points inside the bump region, and wherein the plurality of intensities comprises intensities of the points outside and around the bump region and intensities of the points inside the bump region;
• obtaining fitted function values by fitting a third order polynomial to the intensities of the points outside and around the bump region;
• replacing the intensities of the points inside the bump region with the fitted function values;
• and
• projecting the fitted function values inside the bump region on the 2D projection image back to the 3D VOI.

A system of local surface smoothing of a three-dimensional (“3D”) binary volume of interest (“VOI”), wherein the VOI comprises a plurality of 3D surface points, the system comprising:

• means for identifying a projection plane;
• means for obtaining a two-dimensional (“2D”) projection image by projecting the plurality of 3D surface points onto the projection plane;
• means for determining a plurality of intensities of the projection image, the plurality of intensities comprising distances from the plurality of 3D surface points to the projection plane;
• means for identifying a bump region on the 2D projection image, wherein the plurality of 3D surface points comprises points outside and around the bump region and points inside the bump region, and wherein the plurality of intensities comprises intensities of the points outside and around the bump region and intensities of the points inside the bump region;
• means for obtaining fitted function values by fitting a third order polynomial to the intensities of the points outside and around the bump region;
• means for replacing the intensities of the points inside the bump region with the fitted function values;
• and
• means for projecting the fitted function values inside the bump region on the 2D projection image back to the 3D VOI.