This folder contains the demo workflow described in our paper, Evaluating the resolution of AI-based accelerated MR reconstruction using a deep learning-based model observer.

The demos are organized so that each stage has a clear role in the workflow:

1. Demo 1: Object generation using DDPM

   Generate DDPM-based MR object samples and save them as .npz files.

2. Demo 2: Synthetic defect insertion

   Insert singlet and doublet defects into the generated objects.

3. Demo 3: MR acquisition and reconstruction

   Simulate accelerated acquisition and perform rSOS reconstructions.

4. Demo 4: DLMO training

   Train the model observer with prepared training and validation datasets.

5. Demo 5: A simple example of the DLMO framework

   Run a compact end-to-end example using bundled sample objects and provided checkpoints.

6. Demo 6: Statistical analysis

   Run the sample size estimation and pivotal-study scripts used in the paper.

Please refer to each demo README for prerequisites, inputs, outputs, and example commands.