BioMed Workspace

A web-based platform for biomedical image processing that brings advanced machine learning to researchers without requiring programming expertise.

Launch Workspace

What Can You Do?

The Biomedical Image Processing Workspace provides a complete pipeline from raw microscopy data to publication-ready visualizations.

Process Microscopy Images

Whether you work with confocal microscopy, light-sheet imaging, or electron microscopy data, the workspace handles the computational heavy lifting.

Train Deep Learning Models

Train U-Net segmentation models and self-supervised denoising models (auto-routed autoStructN2V / Noise2Void) on your specific samples - no coding required. The step-by-step workflows guide you through each stage.

Generate 3D Meshes

Convert segmented image stacks into 3D surface meshes for visualization, analysis, or 3D printing. Output formats include OBJ, STL, and Three.js JSON.

Processing Modules

Module	Description
Image Viewer	Browse TIFF image stacks
Preprocessing	Crop, trim, flip/rotate, downscale, and intensity-adjust stacks non-destructively
Deep Learning Denoising	Self-supervised denoising with auto-routed autoStructN2V and N2V
Filter Denoising	Classical Gaussian and Non-Local Means filtering
Quick Annotation	Browser-based annotation tool for ground truth creation
U-Net Segmentation	Train and run semantic segmentation models
Segmentation Cleanup	Paint corrections, automated cleanup, and per-class quantification
Stack Stitching	Join stacks in z or as mosaics, with reusable stitch recipes
Mesh Generation	Convert segmentations to 3D polygon meshes
3D Visualization	Interactive mesh viewer with clipping planes
File Browser	Manage your workspace files and exports

Get Started

Overview

Step-by-step guides and best practices

Getting Started

Welcome to the Biomedical Image Processing Workspace.

Info & Help Panel

Find the information you need directly in the Workspace

For Different Research Approaches

For Practical Imaging Work

The workspace removes the technical barriers to applying machine learning to your microscopy data. Built-in test data lets you explore features before uploading your own samples. Parameters are explained in context, and sensible defaults work for most cases.

For Computational and Quantitative Work

The workspace provides a transparent ML pipeline with full access to training dynamics and results. Monitor loss curves, validation metrics, and training progress in real-time. Results export in standard formats (TIFF, PyTorch, OBJ/STL) compatible with downstream analysis.