Virtual NEMA / IEC Nuclear Medicine Phantom Generator

phantomgen is a lightweight Python package that generates realistic 3D numerical phantoms for Nuclear Medicine imaging simulations and dosimetry research.
It implements the IEC Body Phantom (NEMA NU-2 / IQ) geometry and provides parameter sets matching EARL and PET standard configurations.

• Generate 3D activity maps and CT μ-maps with realistic geometry
• Supports spherical inserts, and lung inserts
• Adjustable voxel size and matrix dimensions
• Ready-made parameter dictionaries for PET and EARL configurations
• Pure NumPy implementation (no heavy dependencies)

The NEMA (National Electrical Manufacturers Association) IQ Body Phantom is a standardised test object used for quality control and performance evaluation in PET and SPECT systems.
phantomgen provides a digital version of this phantom, suitable for:

• Monte Carlo simulations (e.g., GATE, SIMIND, STIR)
• Deep-learning–based dosimetry and image reconstruction
• Algorithm testing and reproducibility studies

# From source
git clone https://github.com/varzakis/phantomgen.git
cd phantomgen
pip install -e .

Requires Python ≥ 3.9 and NumPy ≥ 1.23.

import numpy as np
from phantomgen import create_nema, pet_nema_dict, earl_nema_dict

# Create a PET phantom (256³ voxels, 2 mm voxel size)
act_vol, ct_vol = create_nema(
    matrix_size=(256, 256, 256),
    voxel_size_mm=(2.0, 2.0, 2.0),
    nema_dict=pet_nema_dict
)

np.save("activity.npy", act_vol)
np.save("ctmu.npy", ct_vol)

After installation, run from the terminal:

phantomgen --preset pet --out-act act.npy --out-ct ct.npy

Optional arguments:

Example:

phantomgen --preset earl --z 192 --y 192 --x 192 --voxel 2.5 2.5 2.5

Both are returned as 3D NumPy arrays (Z, Y, X) in float32 format.

Both default configurations (pet_nema_dict, earl_nema_dict) can be customized before generation.

Example:

from phantomgen import create_nema, pet_nema_dict

custom = pet_nema_dict.copy()
custom["sphere_dict"]["spheres"]["act_conc_MBq_ml"] = [1.0]*6
act, ct = create_nema(nema_dict=custom)

A minimal test suite is included. Run it with:

pytest -q

MIT License © 2025 — Your Name
This software is free and open source.

import matplotlib.pyplot as plt
from phantomgen import create_nema, pet_nema_dict

act, ct = create_nema(nema_dict=pet_nema_dict)
z = act.shape[0] // 2

fig, ax = plt.subplots(1, 2, figsize=(10, 5))
ax[0].imshow(act[z], cmap="inferno")
ax[0].set_title("Activity (MBq/voxel)")
ax[1].imshow(ct[z], cmap="gray")
ax[1].set_title("CT μ-map (cm⁻¹)")
plt.show()

phantomgen/
├─ src/phantomgen/core.py       # Core implementation
├─ src/phantomgen/__init__.py   # Public API
├─ tests/test_basic.py          # Minimal tests
├─ pyproject.toml
└─ README.md

Enjoy generating your virtual NEMA phantoms 💫