Information processing method, medical image diagnostic apparatus, and information processing system including deep learning for optimal cardiac phase selection

A method, comprising:

• receiving a phase image acquired at a corresponding cardiac phase;
• determining, from the received phase image, a mask image of a particular cardiac region;
• applying both the determined mask image and the phase image to inputs of a trained neural network model to obtain, from outputs of the trained neural network model, a location probability map, the trained neural network model having been trained with a set of input training data and a corresponding set of output training data, wherein the input training data includes a training mask image and a training phase image, and the output training data includes a training location probability map;
• and
• calculating, for the cardiac phase, from the determined location probability map output from the trained neural network model, a value of a cardiac motion metric, wherein the determined location probability map specifies a probable location of a cardiac vessel.

A non-transitory computer-readable medium storing a program that, when executed by processing circuitry, causes the processing circuitry to perform a method, comprising:

• receiving a phase image acquired at a corresponding cardiac phase;
• determining, from the received phase image, a mask image of a particular cardiac region;
• applying both the determined mask image and the phase image to inputs of a trained neural network model to obtain, from outputs of the trained neural network model, a location probability map, the trained neural network model having been trained with a set of input training data and a corresponding set of output training data, wherein the input training data includes a training mask image and a training phase image, and the output training data includes a training location probability map;
• and
• calculating, for the cardiac phase, from the determined location probability map output from the trained neural network model, a value of a cardiac motion metric, wherein the determined location probability map specifies a probable location of a cardiac vessel.