Works with your entire QuPath stack

Instance segmentation for nuclei and cells using deep learning. Steps are fully captured — but model files must be downloaded on the replay machine before running the script.

Star-convex polygon detection for nuclei. Widely validated on H&E and fluorescence. Supports pre-trained and custom models.

Patch-based deep learning inference for whole-slide images. Tile-level classification and scoring from the WSInfer model hub.

Run any model from the BioImage Model Zoo within QuPath — hundreds of community-trained segmentation and detection models.

Deep Java Library integration — general-purpose deep learning inference backend used by InstanSeg and other AI extensions.

TensorFlow SavedModel inference for pixel classification and object detection. Legacy backend for older saved models.

Cell and nucleus segmentation via the Cellpose Python server. Highly accurate across diverse cell morphologies and imaging modalities.

Interactive segmentation using Meta AI's Segment Anything Model. Brush-guided and point-prompted annotation in QuPath.

Read 150+ proprietary microscopy formats: CZI, LIF, ND2, VSI, SVS, and more. The de-facto standard for multi-dimensional image I/O in life sciences.

High-performance reading of whole-slide image formats including SVS, TIFF, SCN, BIF, and NDPI. Bundled with QuPath.

Browse, open, and annotate images stored on an OMERO server. Bidirectional synchronisation of ROIs and metadata.

Export annotation overlays and image regions as scalable SVG files. Essential for publication-quality figures.

Non-rigid multi-modal image registration using BigWarp. Align serial sections, multi-stain panels, or DESI-MS with H&E.

Overlay and combine multiple registered image channels using Warpy transformations. Designed for multiplexed and multi-modal panel workflows.

Utilities and processing helpers from the BioImaging & Optics Platform at EPFL. Measurement tools, batch processing helpers, and workflow utilities.

Run ImageJ/Fiji macros on QuPath image regions. Access the full ImageJ plugin ecosystem from within a QuPath script.

Built-in pixel and object classifier training: Random Forest, ANN, SVM. Classifiers trained in the QuPath UI are fully captured and embedded in the .qtrace.

Core analysis tools: cell detection, tissue segmentation, subcellular spot detection, and measurement-based classification. Fully scriptable and captured.