machine learning for image interpolation in CT scan

Hi, I have setup a CT scan, now I want to use machine learning to interpolate the images in order to reduce the number of projections. It is limited angle CT system, in case I have had 2D projected image at 0,20,40,60,80,100,120,140,160,and 180 degrees, now I want to generate a image sequence from 0 to 180 degree with 1 degree angular space. which leaning method or structure is recommanded? Anybody has similar experience?

Hi @qdlcwk,

It sounds to me like you are trying to solve a very particular problem here. So please don’t expect a readily available solution.

For clarification: you already have 2D slices/images at the mentioned angles reconstructed?

And subsequently you want to use the interpolated sequence for 3D reconstruction? If so, I would assume that you’d get better results with a more sophisticated method that improves the 3D reconstruction directly from the raw input images (e.g. [1711.10388] Lose The Views: Limited Angle CT Reconstruction via Implicit Sinogram Completion). But please take this with a grain of salt since I am no expert in medical imaging.

I assume that it would be possible to use CARE if you have enough training data with the full resolution (1 degree steps) that you can downsample to create the input for the CNN. You could, in addition, contact the authors with a more detailed description of what you are trying to achieve…



This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.