HIPPOCAMPUS EXTRACTION METHOD AND SYSTEM USING A GRAPH-CUTS ALGORITHM COMBINED WITH ATLAS-BASED SEGMENTATION AND MORPHOLOGICAL OPENING

摘要

Disclosed are a method and a system for extracting a hippocampus by assembling a map-based segmentation and a morphological computation and applying to a graph-cut algorithm. The method for extracting a hippocampus area from an input image includes the following steps of: matching the input image with a template image; generating a partial volume image about a gray matter tissue, a white matter tissue, and a brain spinal cord based on the matched input image; matching the matched input image with the map of the hippocampus in a nonlinear way and separating the image of the hippocampus area from the nonlinearly matched input image; generating a foreground image and a background image about the image of the hippocampus area based on the image of the hippocampus area which is separated from the generated partial volume image and the matched input image; and extracting the hippocampus area using the graph-cut algorithm based on the generated foreground and background images. [Reference numerals] (AA) Start; (BB) End; (S110) Match an input image with a template image; (S120) Generate PVE maps using an ANN algorithm based on the matched image; (S130) Separate a hippocampus area from the matched image using a map-based segmentation method; (S140) Generate a foreground image and a background image based on the PVE maps and the map-based segmentation result; (S150) Extract the hippocampus area using a graph-cut algorithm based on the foreground and background images