Vol. 0.1 · No. 7

stainkit

GPU-accelerated H&E stain normalisation · A field deployable

Abstract.

stainkit implements the Macenko, Ruifrok–Johnston, and Otsu algorithms in a single C++/CUDA binary, exposing them as a CLI, a static library, and a Python module. Upload an H&E patch and receive a stain-normalised image, an Otsu tissue mask, and a side-by-side comparison in under a second on a single Tesla T4.

Fig. 1 · Sample inputs.

Six synthetic H&E patches spanning a range of stain characteristics. Click any to load.

Methods · Upload.

PNG, JPEG, BMP, or TGA. Maximum 10 MB.

Idle.

References.

  1. Macenko M, Niethammer M, Marron JS, Borland D, Woosley JT, Guan X, Schmitt C, Thomas NE. A reference image set for H&E stain normalisation. Proc IEEE Int Symp Biomed Imaging. 2009; 233–236. doi:10.1109/ISBI.2009.5193250
  2. Ruifrok AC, Johnston DA. Quantification of histochemical staining by color deconvolution. Anal Quant Cytol Histol. 2001;23(4):291–299. PMID 11531144
  3. Otsu N. A threshold selection method from gray-level histograms. IEEE Trans Syst Man Cybern. 1979;9(1):62–66. doi:10.1109/TSMC.1979.4310076