pept-forward-model

Positron Emission Particle Tracking (PEPT) is a nuclear imaging technique for observing the movement of particles in opaque systems.

This is a very simplistic 'forward' model for simulating data observed by a real world detector. It is simple to use and understand compared to the more sophisticated software available and models two main sources of uncertainty in the problem:

• Compton scattering of the photons.
• The discrete size of the cells that make up the detector.

This is primarily research code, so expect bugs.

Usage

First it is necessary to create the relevant objects.

from model import StaticParticle, CylinderDetector

p = StaticParticle()
d = CylinderDetector()

This creates a cylindrical detector and a particle at the origin tagged with a radioactive tracer. The default parameters are suitable for most cases, however we can configure certain aspects when needed

# Move particle closer to the detector and up
p.set_position_cylindrical(r=0.1, theta=0, z=0.1)

# Adjusts the size of the detector cells in cm
d.detectors_height = 0.01
d.detectors_width = 0.01

To generate Line of Response (LoR) data, there is a single function call

lors, scatters = p.simulate_emissions(detector=d, n_emissions=10**4)

It is encouraged to examine the source code for these classes to familiarise oneself with the inner workings and model parameters that can be set. Example code is available (see experiments.ipynb) for help plotting and analysing the emission data.

Note that the scripts in this repository are intended to be run as modules from the source root, e.g

python -m inversion/generate_plots.py

This software is largely research code so expect bugs.

Installation

pip install -r requirements.txt