Automatic spinal canal segmentation using cascaded random walks

An apparatus for spinal canal segmentation in a 3D medical image, comprising:

• means for detecting an initial set of spinal canal voxels in the 3D medical image using a trained classifier;
• means for refining a spinal canal topology defined by a current set of spinal canal voxels based on an estimated medial line of the spinal canal;
• means for sampling seed points based on the refined spinal canal topology;
• and
• means for iteratively detecting updated sets of spinal canal voxels in the 3D medical image using cascaded random walks segmentation based on the sampled seed points.

A non-transitory computer readable medium storing computer program instructions for spinal canal segmentation in a 3D medical image, the computer program instructions when executed by a processor cause the processor to perform operations comprising:

• detecting an initial set of spinal canal voxels in the 3D medical image using a trained classifier;
• refining a spinal canal topology defined by a current set of spinal canal voxels based on an estimated medial line of the spinal canal;
• sampling seed points based on the refined spinal canal topology;
• and
• detecting an updated set of spinal canal voxels in the 3D medical image using random walks segmentation based on the sampled seed points.