The Histology
Platform

Combining annotation tools, deep learning, and quantitative analysis into an engine of discovery.

Platform Overview

Our platform enables precise annotation of histological structures, training of state-of-the-art segmentation models, and quantitative analysis — generating reproducible, publication-ready results with full provenance.

We built precision tools for sub-pixel boundary tracing.

Our annotation editor provides polygon and freehand brush tools with real-time pan, zoom, and vertex editing — designed for the demands of high-resolution microscopy images where accuracy at the cellular level matters.

Multi-class ROI support with distinct color coding, auto-save, and mask import from external tools (indexed or RGB with palette mapping) means you can integrate HistoScope into existing workflows without friction.

Polygon & freehand tools

Click-to-place vertices or draw freehand with RDP simplification.

Multi-class ROI support

Define unlimited classes with distinct colors and label values.

Mask import & validation

Import indexed or RGB masks from QuPath, ImageJ, or other tools.

R&D Workflow

Platform-powered
discovery

01

Upload & organize

Create projects, define ROI classes, upload histology images in PNG/JPEG/TIFF. Organize into training, validation, and test datasets.

  • Multi-format support (8-bit and 16-bit)
  • Bulk upload with metadata tracking
  • Import pre-annotated masks from other tools
02

Annotate with precision

Open images in the annotation editor. Use polygon or freehand tools to label regions of interest. Changes auto-save as you work.

  • Sub-pixel accuracy at any zoom level
  • Multi-class support with distinct colors
  • Keyboard shortcuts for efficient workflow
03

Train & validate

Select datasets, configure model architecture and augmentation pipeline. Monitor training in real-time with live loss curves.

  • UNet with configurable encoders
  • Three presets or full customization
  • Run comparison and versioning
04

Analyze & export

Run inference, review predictions, measure ROIs, and export results. Every output includes full provenance — model, data, parameters, timestamps.

  • Batch inference processing
  • ROI measurements (area, perimeter)
  • Export as PNG, CSV, or JSON
Our Approach

Built for reproducible research, not just pretty pictures.

Every design decision in HistoScope prioritizes correctness, traceability, and reducing friction. Three principles guide everything we build.

PRECISION

Correctness Over Features

Wrong measurements or masks mean product failure. Every coordinate is in image-space, every mask is validated, every metric is verifiable.

PROVENANCE

Trust & Traceability

Every result shows the model, dataset, parameters, and timestamps that produced it. Full audit trail from annotation to export.

EFFICIENCY

Reduce Friction

Fewer clicks, prevent mistakes, clear next steps. Auto-save, keyboard shortcuts, batch operations, and smart defaults throughout.

Our Focus

Histology segmentation for research labs.

Multi-tenant

PI-led labs with role-based access control. Students request training, PIs approve. Invite members with shareable links.

Full Pipeline

From raw images to exported segmentation results. Annotation, training, inference, and analysis in a single platform.

Open Stack

Built with PyTorch, FastAPI, Next.js, and PostgreSQL. Docker Compose for local deployment. S3-compatible storage.

We are building precision tools for histology segmentation research.

Get Started