Image Reconstruction using Artificial Intelligence (Ai) Modules Agnostic to Image Acquisition Settings

An image processing system for determining an image of a scene, comprising:

• an input interface configured to accept measurements of the scene acquired under image acquisition settings and settings data indicative of the image acquisition settings;
• a memory configured to store instructions implementing modules of the system, the modules comprising:
• an image acquisition settings estimator module configured to estimate a forward operator of the image acquisition settings based on the settings data;
• an image acquisition inverter module configured to modify a previous estimate of the image of the scene to fit the measurements inverted with the forward operator to produce a current noisy image of the scene;
• a denoiser module configured to remove noise from a noisy image provided to the denoiser module subject to a known noise level;
• a noise level estimator module configured to estimate a noise level in an input image provided to the noise level estimator module;
• a processor coupled with the memory and configured to
• execute the image acquisition settings estimator module to estimate the forward operator using the setting data;
• and
• recursively estimate the image of the scene until a termination condition is met, wherein for a current iteration, the processor (1) executes the image acquisition inverter module with the estimated forward operator and the previous estimate of the image of the scene estimated during a previous iteration to produce the current noisy image;
• (2) executes the noise level estimator module with the previous estimate of the image to produce a current noise level;
• and (3) executes the denoiser to remove the current noise level from the current noisy image to produce a current image;
• and
• an output interface configured to output the image of the scene.