Low-dimensional manifold constrained disentanglement network for metal artifact reduction

A method for metal artifact reduction (MAR) in computed tomography (CT) images, the method comprising:

• constructing, by a patch set construction module, a patch set based, at least in part on training data;
• determining, by a manifold dimensionality module, a dimensionality of a manifold related to the patch set;
• and
• optimizing, by a training module, a combination loss function, the combination loss function comprising a network loss function and the manifold dimensionality, wherein optimizing the combination loss includes simultaneously optimizing the network loss function and the manifold dimensionality, and wherein optimizing the combination loss function comprises optimizing at least one network parameter;
• wherein the training data comprises paired images and unpaired images, the paired images corresponding to synthesized paired data, and the unpaired images corresponding to unpaired clinical data.

A system for metal artifact reduction (MAR) in computed tomography (CT) images, the system comprising:

• a computing device comprising a processor, a memory, an input/output circuitry, and a data store;
• a patch set construction module configured to construct a patch set based, at least in part on training data;
• a manifold dimensionality module configured to determine a dimensionality of a manifold related to the patch set;
• and
• a training module configured to optimize a combination loss function, the combination loss function comprising a network loss function and the manifold dimensionality, wherein optimizing the combination loss function includes simultaneously optimizing the network loss function and the manifold dimensionality, and wherein optimizing the combination loss function comprises optimizing at least one network parameter;
• wherein the training data comprises paired images and unpaired images, the paired images corresponding to synthesized paired data, and the unpaired images corresponding to unpaired clinical data.