Three-Dimensional Shape Reconstruction from a Topogram in Medical Imaging

A method for machine-training to reconstruct a three-dimensional representation of anatomy from a first topogram in a medical imaging system, the method comprising:

• defining a machine learning architecture having a first encoder configured to receive second topograms as input and a second encoder configured to receive masks or first three-dimensional shapes as input;
• machine learning with the machine learning architecture to output the three-dimensional representation in response to input of the first topogram;
• and
• storing a machine-learned model resulting from the machine learning.

A medical imaging system for three-dimensional organ estimation, the medical imaging system comprising:

• an X-ray imager to generate an X-ray image of a patient;
• an image processor configured to generate an estimation of the three-dimensional organ from input of the X-ray image to a machine-learned model, the machine-learned model configured to receive as input views from a single direction, the X-ray image being a view from the single direction;
• and
• a display configured to display information from the estimation of the three-dimensional organ.