Method and System for Deep Motion Model Learning in Medical Images

An apparatus for motion estimation and modeling in a medical image sequence of a patient, comprising:

• means for receiving a medical image sequence of a patient;
• means for inputting a plurality of frames of the medical image sequence to a trained deep neural network;
• means for generating, using the trained deep neural network, diffeomorphic deformation fields representing estimated motion between the frames of the medical image sequence input to the trained deep neural network;
• and
• means for generating an encoding of observed motion in the medical image sequence using the trained deep neural network;
• means for predicting non-observed motion from the medical image sequence and generating at least one predicted frame using the trained deep neural network, wherein the at least predicted frame is one of a predicted future frame or a predicted frame between the frames of the medical image sequence input to the trained deep neural network.

A non-transitory computer readable medium storing computer readable instructions for computer-based motion estimation and modeling in a medical image sequence of a patient, the computer program instructions when executed by a processor cause the processor to perform operations comprising:

• receiving a medical image sequence of a patient;
• inputting a plurality of frames of the medical image sequence to a trained deep neural network;
• generating, using the trained deep neural network, diffeomorphic deformation fields representing estimated motion between the frames of the medical image sequence input to the trained deep neural network;
• and
• generating an encoding of observed motion in the medical image sequence using the trained deep neural network;
• predicting non-observed motion from the medical image sequence and generating at least one predicted frame using the trained deep neural network, wherein the at least predicted frame is one of a predicted future frame or a predicted frame between the frames of the medical image sequence input to the trained deep neural network.