Systems and methods for computed tomography image reconstruction

A method, comprising:

• obtaining a first computed tomography (CT) image, the first CT image reconstructed from a high helical pitch dataset;
• inputting the first CT image to a trained multidimensional statistical regression model trained to output a second CT image based on the first CT image, the trained multidimensional statistical regression model trained to map the first CT image to the second CT image using a training dataset that includes a plurality of training CT image pairs, each training CT image pair including a first training CT image reconstructed from a low helical pitch dataset and a second training CT image reconstructed from a pseudo high helical pitch dataset generated from the low helical pitch dataset;
• iteratively correcting the second CT image based on the first CT image to generate a final CT image;
• and
• storing the final CT image and/or outputting the final CT image for display.

An image processing system comprising:

• a display device;
• a user input device;
• a trained multi-dimensional statistical regression model;
• a processor communicably coupled to the display device, the user input device, and a non-transitory memory storing the trained multidimensional statistical regression model and including instructions that when executed cause the processor to:
• in a first step, reconstruct a first computed tomography (CT) image of a region of interest (ROI) of a subject from high helical pitch projection data, using interpolated data to fill in missing projection data of the high helical pitch projection data;
• in a second step, generate a second CT image using the trained multidimensional statistical regression model, the second CT image of higher quality than the first CT image;
• in a third step, iteratively correct the second CT image based on the first CT image;
• and
• display the second CT image via the display device;
• and wherein the trained multidimensional statistical regression model has a neural network architecture.