This invention belongs to the field of medical
image processing technology, specifically relating to a
kidney segmentation method based on key
point localization. The method includes the following steps: acquiring medical images of the
kidney to be segmented and preprocessing the images; defining a set of anatomical key points specifically for
kidney segmentation; predicting the position coordinates of each key point in the key
point set using a model; constructing spatial topological relationships between key points based on the predicted key point coordinates and kidney anatomical
zoning rules; and dividing the kidney
parenchyma into regions according to the spatial topological relationships and kidney anatomical
zoning rules, outputting the segmentation results for each kidney segment. This invention achieves efficient and accurate automatic kidney segmentation by precisely detecting anatomical key points and combining them with the anatomical rules of the Graves kidney segment model, outputting segmentation results with anatomical
interpretability, and providing reliable
technical support for
preoperative planning of urological surgeries such as precise nephrectomy.