High-Resolution 4D CT Reconstruction

The monai-physio-reconstruct-highres-4d-ct command reconstructs a high-resolution 4D CT time series from ordered phase images and a high-resolution reference image.

Input Requirements

  • Ordered 3D phase images, such as .mha, .mhd, .nrrd, or .nii.gz files.

  • A fixed high-resolution reference image.

  • Optional fixed and moving masks for registration focus.

TCIA-4DLung’s tutorial subset downloads automatically via monai-physio-download-data TCIA-4DLung; the full TCIA 4D-Lung collection is still obtained manually - downloading its DICOM series from TCIA and running data/TCIA-4DLung/convert.py - if more than the tutorial subset is needed. See data/TCIA-4DLung/README.md for the download and layout instructions.

Basic Usage

monai-physio-reconstruct-highres-4d-ct \
    --time-series-images frame_*.mha \
    --fixed-image highres_reference.mha \
    --output-dir ./results

Choosing a Reference Frame

monai-physio-reconstruct-highres-4d-ct \
    --time-series-images frame_000.mha frame_001.mha frame_002.mha \
    --fixed-image highres_reference.mha \
    --reference-frame 0 \
    --output-dir ./results

Registration Options

monai-physio-reconstruct-highres-4d-ct \
    --time-series-images frame_*.mha \
    --fixed-image highres_reference.mha \
    --registration-method Greedy_ICON \
    --Greedy-iterations 30 15 7 3 \
    --prior-weight 0.5 \
    --output-dir ./results

Composite Mode

By default, the high-resolution reference image is warped back to each time point. --composite-mode mean or --composite-mode max instead build a pixel-by-pixel mean or max composite of the reference image and every registered time-series image first, then warp that composite back to each time point - useful when anatomy or contrast is only visible in some frames.

monai-physio-reconstruct-highres-4d-ct \
    --time-series-images frame_*.mha \
    --fixed-image highres_reference.mha \
    --composite-mode mean \
    --output-dir ./results

Outputs

The command writes reconstructed images to --output-dir using --output-prefix as the filename prefix. Use --save-transforms and --save-losses when registration diagnostics are needed.

Python API

Use monai_physio.WorkflowReconstructHighres4DCT for programmatic access.