VI-PRISM: Perfusion Imaging and Single Material Reconstruction in Polychromatic Photon Counting CT

This repository implements and evaluates a VI-based reconstruction method for dose-reduced perfusion CT in polychromatic photon-counting CT. The method adapts a theoretically motivated monotone variational inequality (VI) algorithm to the perfusion setting, where the static background tissue is assumed known and the goal is to reconstruct the iodine (contrast agent) concentration map.

Experiments are run on a digital phantom with water and iodine at varying concentrations. The code sweeps key acquisition parameters to study dose and sampling trade-offs:

• Iodine concentration: 0.05 to 2.5 mg/ml
• Photon budget (mean photons per detector element): 1e5 down to 1e2
• Number of projections: 984 down to 8

Environment setup (conda)

conda env create -f environment.yml

Quick start

To run the full experiment suite, execute:

python runner.py