Reading images#

imread() is the single entry point for getting data into DarSIA. It looks at the file extension, dispatches to the right reader, and returns an Image (or a list of images for a folder or a list of paths).

import darsia

# one optical photograph, 2.8 m x 1.5 m
photo = darsia.imread("baseline.jpg", width=2.8, height=1.5)

# a whole folder of photographs as one space-time image
series = darsia.imread_from_optical(["f/img_0.jpg", "f/img_1.jpg"], dimensions=[1.5, 2.8])

# a 3D DICOM stack
volume = darsia.imread("scan/", dim=3)

Supported formats#

optical (jpg/png/tif)	`imread_from_optical()` – returns `OpticalImage`
DICOM (.dcm)	`imread_from_dicom()` – 2D/3D/4D medical stacks
simulation (.vtu)	`imread_from_vtu()` – meshio-backed, mixed-dimensional
NumPy (.npy/.npz)	`imread_from_numpy()`, `imread_from_npz()`
in-memory bytes	`imread_from_bytes()`

Applying corrections while reading#

Pass correction objects via transformations= and they run as the image is read, so downstream code only ever sees corrected data:

curvature = darsia.CurvatureCorrection(config=cfg["curvature"])
photo = darsia.imread("baseline.jpg", transformations=[curvature],
                      width=2.8, height=1.5)

See Shape correction, Colour correction and Illumination correction for the correction catalogue.

See also#