Probabilistic segmentation of volumetric images

1. A computer-implemented method for automated segmentation of volumetric images, said method comprising the steps of:

receiving a coarse volumetric image comprising at least one anatomical structure in terms of voxels;

defining each voxel a distinct anatomical identifier based on a probabilistic distribution for each of the anatomical structure;

applying the defined coarse volumetric image through a coarse model for a coarse output;

combining the coarse output with the coarse volumetric image;

generating a probabilistic segmentation from applying the combined coarse output/image through a fine model; and

converting the probabilistic segmentation to a polygonal mesh for each defined class applying a volume-to-mesh algorithm for at least one of a dental anatomy condition, or treatment.

2. The method of claim 1 , further comprising a patch-based approach for the fine model segmentation.

3. The method of claim 1 , further comprising a Region-of-Interest (RoI) approach for the fine model segmentation.

4. The method of claim 1 , wherein the at least one received coarse volumetric image comprises a 3-D pixel array.

5. The method of claim 4 , further comprising the step of pre-processing by converting the 3-D pixel array into an array of Hounsfield Unit (HU) radio intensity measurements.

6. The method of claim 1 , further comprising the step of pre-processing the received coarse volumetric image by rescaling using linear interpolation.

7. The method of claim 6 , wherein the pre-processing comprises using any one of a normalization schemes to account for variations in image value intensity depending on at least one of an input or output of the coarse volumetric image.

8. The method of claim 1 , further comprising the step of localizing using a V-Net-based fully convolutional neural network.

9. The method of claim 1 , further comprising the step of extracting anatomical structure by finding a minimum bounding rectangle around the defined structure.

10. The method of claim 9 , wherein the bounding rectangle extends by at least 15 mm vertically and 8 mm horizontally (equally in all directions) to capture the tooth and surrounding context.

11. The method of claim 1 , further comprising the step of condition classification using a DenseNet 3-D convolutional neural network.

12. A system for automated segmentation of volumetric images, said system comprising:

a processor;

a non-transitory storage element coupled to the processor;

encoded instructions stored in the non-transitory storage element, wherein the encoded instructions when implemented by the processor, configure the system to:

receive a coarse volumetric image comprising at least one anatomical structure in terms of voxels;

define each voxel a distinct anatomical identifier based on a probabilistic distribution for each of the anatomical structure;

apply the defined coarse volumetric image through a coarse model for a coarse output;

combine the coarse output with the coarse volumetric image; and

generate a probabilistic segmentation from applying the combined coarse output/image through a fine model; and

convert the probabilistic segmentation to a polygonal mesh for each defined class applying a volume-to-mesh algorithm for at least one of a dental anatomy, condition, or treatment.

13. The system of claim 12 , further comprising a patch-based approach for the fine model segmentation.

14. The system of claim 12 , further comprising a Region-of-Interest (RoI) approach for the fine model segmentation.