Reinhard

class Reinhard(StainNormalizerBase)

Constructor

Reinhard(
    device: str | torch.device | None = None,
    backend: str | None = None
)

Parameters: - device: Device to use. Default: None (auto-detect) - backend: Backend ("pytorch" or "cuda"). Default: None (auto-select)

Methods

• fit(reference_images) - Compute reference mean/std in LAB color space
• transform(images) - Apply Reinhard normalization

Example

from stainx import Reinhard

normalizer = Reinhard(device="cuda")
normalizer.fit(reference)
normalized = normalizer.transform(images)