Knowledge-based automatic image segmentation

A method for segmenting new image data sets, comprising:

• receiving a medical image set;
• searching the knowledge-base for representative matches to form a base set, wherein the base set comprises a plurality of reference image sets;
• deforming a portion of the plurality of reference image sets of the base set to generate mappings from the base set to the medical image set;
• determining a first weighted-average segmentation for a plurality of structures of interest in the medical image set,
• searching the knowledge-base for representative matches to form a plurality of sub-volume base sets, wherein a sub-volume base set comprises a plurality of reference image set sub-volumes;
• deforming a portion of the plurality of reference image set sub-volumes of the plurality of sub-volume base sets to generate mappings from the plurality of sub-volume base sets to corresponding structures of interest of the medical image set;
• and
• determining a second weighted-average segmentation for the plurality of structures of interest in the medical image set.

A computer readable storage media having computer-readable and computer-executable program code embodied therein for causing a computer system to execute a method of segmenting new image data sets, comprising:

• receiving a medical image set;
• computing a sparse landmark signature based on the medical image set, wherein the sparse landmark signature comprises a plurality of points;
• searching a knowledge-base for representative matches to form a base set, wherein the base set comprises a plurality of reference image sets, and wherein the base set is formed from a cohort of reference image sets within the knowledge-base that share common characteristics, wherein the common characteristics comprise at least one of:
• sex, diagnosis, stage, responder versus non-responders, age, treatment site, and size;
• deforming a portion of the plurality of reference image sets of the base set to generate mappings from the base set to the medical image set;
• determining a weighted-average segmentation for a plurality of structures of interest in the medical image set;
• using the weighted-average segmentation for the plurality of structures of interest as a seed for at least one structure-specific segmentation algorithm;
• updating a plurality of statistical analysis tools that determine the weighted-average segmentation;
• reviewing and correcting the contours;
• and
• adding the approved medical image set and corresponding structures of interest to the knowledge-base as a new reference image set.