PlaqSeg
Count phage plaques in seconds using YOLO-powered AI segmentation. Simply photograph your plate and get instant, accurate results.
Available for mobile, desktop, and HPC workflows as part of the Phage Collection Project at the University of Southampton.
PlaqSeg
AI Plaque Counter
Run PlaqSeg Anywhere
Choose PlaqSegDesktop for local point-and-click workflows or PlaqSegHPC for tiled batch inference on high-resolution plates.
PlaqSegDesktop
Native desktop app for routine plaque counting on standard laptops and workstations.
PlaqSegHPC
Tiled YOLOv2.6-seg pipeline for large microscopy images on HPC clusters and laptops.
pip install plaqseg --index-url https://git.soton.ac.uk/api/v4/projects/14521/packages/pypi/simplePlaqSeg Manual
Open the online manual or download it for offline use.
How It Works
Photograph
Take a photo of your plate, or select an image from your camera roll.
AI Analysis
Our YOLO v26 model segments and counts every plaque in under a second.
Results
Get plaque counts, annotated images, and export-ready data instantly.
PlaqSegHPC Installation
From GitLab Package Registry
pip install plaqseg --index-url https://git.soton.ac.uk/api/v4/projects/14521/packages/pypi/simpleFrom Source (Development)
git clone https://git.soton.ac.uk/phage-collection-project/PlaqSeg-HPC.gitcd PlaqSeg-HPCmkdir -p src/plaqseg/modelscp models/*.pt models/*.tflite src/plaqseg/models/pip install -e ".[dev]"Optional Extras
| Extra | Command | Description |
|---|---|---|
| [gpu] | pip install plaqseg[gpu] | CUDA-enabled PyTorch |
| [tflite] | pip install plaqseg[tflite] | Lightweight TFLite runtime |
| [dev] | pip install plaqseg[dev] | pytest, build, twine |
Leo Skingley
Full Stack Developer — University of Southampton
PlaqSeg was developed as part of the Phage Collection Project at the University of Southampton. The application uses state-of-the-art YOLO v26 models and CNNs for instance segmentation to detect and count bacteriophage plaques on agar plates, accelerating phage research workflows.