Methods and apparatus for neural network based image reconstruction

A non-transitory computer readable medium storing instructions that, when executed by at least one processor, cause the at least one processor to perform operations comprising:

• receiving sinogram data from an image scanning system;
• applying a plurality of masks to the sinogram data to generate a plurality of masked sinogram portions;
• applying a plurality of fully connected layers of a neural network to the plurality of masked sinogram portions, wherein an output of the neural network is a plurality of image patches;
• combining the plurality of image patches to generate a first image,
• receiving a plurality of attenuation maps from the image scanning system, wherein the plurality of attenuation maps correspond to the sinogram data;
• and
• applying a plurality of convolutional layers to the first image and the plurality of attenuation maps to generate a second image.

A system comprising:

• a database;
• and
• at least one processor communicatively coupled to the database and configured to:
• receive sinogram data from an image scanning system;
• apply a plurality of masks to the sinogram data to generate a plurality of masked sinogram portions;
• apply a plurality of fully connected layers of a neural network to the plurality of masked sinogram portions, wherein an output of the neural network is a plurality of image patches;
• combine the plurality of image patches to generate a first image;
• receiving a plurality of attenuation maps from the image scanning system, wherein the plurality of attenuation maps correspond to the sinogram data;
• and
• applying a plurality of convolutional layers to the first image and the plurality of attenuation maps to generate a second image.