Cosmael Thinklab
Cosimaging
A free, all-in-one software for fluorescence microscopy image analysis, with deep specialisation in super-resolution microscopy (SMLM).
What is Cosimaging?
Cosimaging is a free, standalone desktop software for analysing fluorescence microscopy images, with a specialisation in super-resolution microscopy (SMLM). It provides a complete pipeline from raw camera images to localisations data.
It is built for researchers who need to process and analyse SMLM data without switching between multiple tools. No coding, no installation, no licence keys. Download a single executable, drag in your data, and start analysing.
Who is it for?
PhDs & Postdocs
Doing SMLM research with techniques like dSTORM, PALM, or DNA-PAINT.
Core Facilities
Microscopy core facility staff who process data for multiple users.
Biologists
Doing standard widefield or confocal imaging, co-localisation, intensity analysis.
Anyone
Looking for a free, all-in-one alternative to commercial microscopy software.
Why Cosimaging?
All-in-One
Localisation, filtering, drift correction, clustering, co-localisation, resolution, and 3D visualisation in a single app. No more juggling ImageJ + ThunderSTORM + MATLAB + Python scripts.
Publication-Standard Math
Implements peer-reviewed methods: Thompson-Larson-Webb precision, Costes auto-thresholding, Manders coefficients, Fourier Ring Correlation, DBSCAN. Every metric follows the original publications.
No Coding Required
Full graphical interface. Point-and-click from raw data to PDF reports. No scripts, no command line, no Python knowledge needed.
Standalone
A single .exe file. No installation, no dependencies, no licence key. Download and run.
Automated Pipeline
One-click Auto-Analysis wizard runs the entire SMLM pipeline with smart defaults and generates a comprehensive PDF + CSV report.
Batch Processing
Process hundreds of datasets overnight using saved workflow parameters. Reproducible, shareable, scriptless.
Core Capabilities
Everything you need to take fluorescence microscopy data from raw frames to published results.
🔬 Super-Resolution Microscopy (SMLM) Pipeline
A complete workflow for single-molecule localisation microscopy, from raw blinking movies to publication-ready cluster statistics.
- Molecule Localisation — Built-in localiser converts raw TIFF blinking movies into precise molecular coordinates using batched 2D Gaussian fitting (linearised weighted least-squares). Outputs X/Y, photon count, background, PSF width, Thompson-formula precision, and SNR. Camera presets included for Hamamatsu, Andor, and other scientific cameras.
- Data Quality Check — Grades localisation data from A (Excellent) to D (Poor) on every metric. Technique-specific thresholds for dSTORM, PALM, and DNA-PAINT. Exportable as a multi-page PDF with histograms and summary tables.
- Localisation Filters — Interactive range sliders for Photons, Background, Precision, PSF Sigma, SNR, and Frame. Auto mode (Gaussian fitting) for smart outlier removal. Real-time 2D view updates. Non-destructive, reset anytime.
- Single Peak Removal — Removes isolated noise spikes using KD-Tree based spatio-temporal nearest-neighbour analysis. Z-axis scaling for 3D data. Shows removal statistics before applying.
- Drift Correction (DME) — Cross-correlation of temporal segments with parabolic sub-pixel refinement. Visual toggle to compare corrected vs original. Multi-layer batch correction supported.
- Image Resolution (FRC) — Fourier Ring Correlation, the gold standard for SMLM resolution measurement. Deterministic odd/even frame split, Tukey windowed FFT, smoothed curve. Reports resolution at the 1/7 threshold (Nieuwenhuizen et al., 2013).
- Cluster Analysis (DBSCAN) — Density-based spatial clustering with separate XY and Z epsilon for 3D data. Ripley’s K analysis to find optimal scale. Reports area, volume, density, radius of gyration, and nearest-neighbour distances. Multi-page PDF output with histograms and per-cluster tables. Save/load parameter presets for reproducibility.
- Positivity Analysis — For multi-channel SMLM, classifies clusters as single-, double-, or triple-positive based on which channels contribute. Spatial merging of cluster centroids across channels.
📐 Widefield & Confocal Image Analysis
Tools that work on standard TIFF microscopy images. No SMLM required.
- Co-localisation — Publication-standard metrics following Manders (1993) and Costes (2004). Costes automatic thresholding separates signal from background without manual tuning. Pearson’s R computed over signal pixels only. Manders’ tM1/tM2 thresholded overlap coefficients. 200-iteration block-scramble significance test (p ≥ 0.95). Cytofluorogram scatter plot. Optional ROI masks.
- Auto Mask (Otsu) — Automatic signal/background separation using Otsu’s thresholding. Creates a binary mask layer for use with co-localisation or clustering.
- Z-Projection (MIP) — Maximum Intensity Projection from multi-frame TIFF stacks, with memory-efficient streaming for large files.
- Intensity Line Profile — Draw a line across an image to measure intensity along the path. Interactive point placement with undo.
🎨 Visualisation & Rendering
- 2D Viewer — High-performance canvas with smooth zoom, pan, and LOD rendering. Multi-layer compositing with customisable LUTs (Green, Red, Blue, Cyan, Magenta, scientific colormaps). Scale bar overlay. CSV layers rendered as scatter or histogram overlays.
- 3D Scatter Viewer — PyVista/VTK-powered, GPU-accelerated, handles millions of points with eye-dome lighting.
- Orthogonal Projections — Simultaneous XY, XZ, YZ views for verifying 3D data quality.
- 3D Volume Projector — Cluster-based density projections with measurement tools, inter-cluster distance measurement, and export.
- Export — Save composite images as high-resolution PNG/TIFF. PDF reports for quality, clusters, and co-localisation. CSV export for all results. JSON workflow logs for reproducibility.
⚡ Automation & Batch Processing
- Auto-Analysis Wizard — 3-step setup (choose mode → select analyses → set parameters). Runs the full pipeline automatically: quality check → filter → drift → FRC → cluster → co-localisation. Generates comprehensive PDF + CSV reports. Supports both SMLM and Widefield modes.
- Batch Processing — Process entire folders of CSV or TIFF files. Replay saved JSON workflows with consistent parameters. Per-dataset PDF, CSV, and log output. Pause, resume, cancel controls.
- Workflow Logs — Every action is timestamped and logged with full parameter details. Save and load workflows as JSON for reproducibility. Shareable between users and sessions.
🧰 Quality of Life
- Layer system — Manage multiple data sources (CSV, TIFF, Cluster, Mask) with visibility toggles, colour pickers, and per-layer properties.
- Drag-and-drop file loading — Just drop your data into the window.
- Column mapping — Auto-detects non-standard CSV column names from ThunderSTORM, rapidSTORM, and others.
- Camera presets — Built-in values for common scientific cameras.
- Keyboard shortcuts — Ctrl+O, Ctrl+S, Ctrl+F, F (fit view), Space (toggle layer), and more.
- Multi-channel TIFF — Load dual-channel side-by-side TIFFs and auto-split into separate layers.
- Beginner guide — Built-in interactive HTML guide for new users.
How It Works
1. Load
Drag and drop your raw TIFF stacks or localisation CSVs. Cosimaging auto-detects column names and camera presets.
2. Analyse
Run the Auto-Analysis wizard for a one-click pipeline, or use individual modules for fine control over every step.
3. Export
Get publication-ready PDF reports, CSV statistics, high-resolution images, and reproducible JSON workflow logs.
Technical Specifications
| Platform | Windows 10 / 11 (standalone .exe) |
| Language | Python 3.10+ |
| GPU | Optional — PyVista/VTK for 3D, CuPy for GPU-accelerated FFT |
| Max dataset size | Tested with 10M+ localisations and 4K×4K TIFF stacks |
| File formats | CSV (any delimiter), TIFF (8/16/32-bit, multi-frame) |
| Dependencies | NumPy, SciPy, Pandas, Matplotlib, scikit-learn, scikit-image, ttkbootstrap |
| Licence | Free |
