daltoscope

See a color, a palette, or a whole image the way a colorblind person does. Where hueristic judges whether colors are safe, daltoscope shows you — it recolors a hex, or an entire PNG/JPEG, as protanopia, deuteranopia, or tritanopia renders it.

pip install daltoscope          # colors only
pip install "daltoscope[image]" # + image recoloring (Pillow, numpy)

import daltoscope as dalton

dalton.simulate("#d7191c", "deutan")   # "#8a7b0c" — that warning red, to a deuteranope
dalton.simulate_all("#1a9641")         # {"protan": "#988839", "deutan": "#8a7f48", "tritan": "#009383"}

from PIL import Image
sim = dalton.simulate_image(Image.open("dashboard.png"), "deutan")
sim.save("dashboard-deutan.png")

API

• simulate(color, type, severity=1.0) → the color as one deficiency sees it, as hex. type is "protan", "deutan", or "tritan".
• simulate_all(color, severity=1.0) → {"protan": ..., "deutan": ..., "tritan": ...}.
• simulate_image(image, type, severity=1.0) → a recolored PIL Image (alpha preserved). image is a PIL Image or a file path. severity (0..1) shows partial deficiency (anomalous trichromacy), not only the full dichromat case.

How it works

daltoscope is a thin lens over opticquiz-cvd: every pixel's transform is that engine's simulate(), using the Machado, Oliveira & Fernandes (2009) deficiency matrices. simulate_image passes each distinct color through the engine once via a lookup table, so a photograph is not thousands of redundant calls — and there is never a second copy of the colorimetry to drift. Method, open access: https://doi.org/10.5281/zenodo.21310578

Not a clinical tool

A simulation for design and communication — it approximates how deficient color vision renders an image. It is not a diagnosis, and no on-screen rendering is exact for a given individual.

Licence

MIT.