Working with Slides¶

The Slide class is the entry point for reading whole slide images.

Slide visualization — Whole slide image: thumbnail, 20x region, and high-magnification region of heart tissue (Aperio AT2)¶

Opening a Slide¶

from glasscut import Slide

slide = Slide("path/to/slide.svs")

By default, Slide attempts to use the CuCim GPU backend first. If CuCim is not available or fails, it falls back to OpenSlide. You can force CPU-only mode:

slide = Slide("path/to/slide.svs", use_cucim=False)

Slide Properties¶

slide.name            # File name without extension
slide.dimensions      # (width, height) at base magnification
slide.magnifications  # List of available magnifications
slide.mpp             # Microns per pixel
slide.properties      # Full metadata dictionary
slide.thumbnail       # PIL Image thumbnail

Context Manager¶

with Slide("slide.svs") as slide:
   print(slide.name)

Backends¶

GlassCut provides two backends implementing the SlideBackend abstract interface:

Backend	Hardware	Dependencies
`OpenSlideBackend`	CPU	openslide-python, openslide-bin
`CuCimBackend`	GPU (CUDA)	cucim, cupy-cuda13x