[ what it does ]
End-to-end analysis
all in one workflow
Automated Cell Analysis Pipeline
OcupathIF takes multi-channel fluorescence microscopy images and runs an end-to-end analysis pipeline — all in a single, guided workflow.
- qptiff format from Akoya Vectra/Polaris systems
- Auto-extracting wavelength metadata
- AI-powered nuclear segmentation with Cellpose 3.x
- Optional SAM 2.1 boundary refinement
- Spectral deblending
- Binary phenotype classification
- Spatial proximity analysis
Interactive Visualization & Review
Researchers can inspect results through an interactive whole-slide viewer. Every step is reviewable and adjustable before export.
- False-color channel overlays
- Adjustable expression gates
- Cell boundary visualization
- Spatial conditioning queries
- “How many CD8+ cells are within 50μm of PD-L1+ cells?”
Built for Speed & Ease of Use
From raw qptiff to CSV-ready results without writing a single line of code.
- Thumbnail-first progressive loading (~9s image preview)
- 10x pipeline speedup over serial methods
- Parallel tile processing
- 5-step guided workflow: View → Analyze → Segment → Count → Export
「 Key Features 」
Automatic channel detection
Reads excitation/emission wavelengths and fluorophore names directly from qptiff metadata
AI cell segmentation
Cellpose 3.x nuclear detection with optional SAM 2.1 boundary refinement
Spectral deblending
Separates overlapping fluorophore signals for accurate phenotyping
Expression gating
Customizable per-channel thresholds with GMM-based automatic gating
Spatial analysis
Proximity queries with configurable distance and visual mask overlays
Multi-language support
English, Chinese, Arabic, Hindi
Cross-platform
macOS (Apple Silicon & Intel) and Windows
One-click install
Standalone app with automatic Python environment setup
[ performance metrics ]
Built for speed at every step
Parallel tile processing and progressive loading make large whole-slide images tractable.
~9s
Image preview load
Thumbnail-first progressive loading
vs ~3 min without
~40ms
Same-image reload
In-memory cache
0.37s
DZI tile generation
Per 5000×5000 channel (pyvips)
10x
Pipeline speedup
Parallel tile processing over serial methods
The fastest path from IF image to insight.
Free, no code required.
macOS & Windows · Free · No code required