Dosimetry Methods

chromadose implements four multichannel film dosimetry methods. All share the same calibration model — the rational function \(\text{pixel}(D) = \frac{r + sD}{t + D}\) — but differ in how they combine information from multiple color channels to estimate dose.

Comparison

Method	Year	Approach	Speed	Best for
Micke	2011	Least-squares multichannel	0.14s	General use, original method
Mayer	2012	Analytical optimization	0.04s	Speed-critical applications
Multigaussian	2018	Maximum likelihood (MLE)	4.7s	Best accuracy, research
ANN	2025	Neural network	~1s	Batch-independent calibration

Times measured on a 550x500 pixel film image.

Unified API

All methods share the same interface:

solver = SomeMethod()
dose_map = solver.solve(film, calibration)

# Every method returns:
dose_map.dose          # (H, W) dose in Gy
dose_map.uncertainty   # (H, W) per-pixel uncertainty
dose_map.dose_r        # (H, W) red channel dose
dose_map.dose_g        # (H, W) green channel dose
dose_map.dose_b        # (H, W) blue channel dose
dose_map.method        # "micke", "mayer", "multigaussian", or "ann"