Cell Detection

Input Formats

Channel handling: cpsam is channel-order invariant — cytoplasm and nuclear channels can be in any order. You do not need to specify which channel is which. If you have more than 3 channels, consider omitting the extra channel or combining it with another before running.

Workflow

1. Load image; detect greyscale vs multi-channel
2. Prepare — pass 1–3 channels through unchanged; truncate >3 to first 3 with a warning
3. Segment with CellposeModel() — no channels argument needed
4. Metrics via skimage.measure.regionprops
5. Figures — overlay + size distribution histogram
6. Report — report.md + {stem}_measurements.csv + reproducibility bundle (commands.sh, environment.yml, checksums.sha256)

CLI Reference

# Standard usage — greyscale or multi-channel (cpsam handles channels automatically)
python skills/cell-detection/cell_detection.py \
  --input <image.tif> --output <report_dir>

# Override diameter estimate (pixels)
python skills/cell-detection/cell_detection.py \
  --input <image.tif> --diameter 30 --output <report_dir>

# Demo (synthetic image, no user file needed)
python skills/cell-detection/cell_detection.py --demo --output /tmp/cell_detection_demo

Demo

python skills/cell-detection/cell_detection.py --demo --output /tmp/cell_detection_demo

Expected output: report.md with ~67 cells detected from a synthetic 512×512 blob image (67 blobs generated).

Algorithm / Methodology

1. Load image with tifffile (TIFF) or PIL (PNG/JPG); detect ndim
2. If >3 channels, truncate to first 3 with a warning
3. Instantiate CellposeModel(gpu=<flag>)
4. Call model.eval(img, diameter=<arg_or_None>) — no channels arg (cpsam is channel-order invariant)
5. Extract per-cell stats from masks via skimage.measure.regionprops
6. Save {stem}_measurements.csv, figures, report.md

Key parameters:

• Model: cpsam (Cellpose 4.0 unified model — channel-order invariant)
• Channels: not passed — cpsam uses the first 3 channels of the input in any order
• Diameter: None triggers Cellpose auto-estimation

Example Queries

• "Segment the cells in my DAPI image"
• "How many cells are in this microscopy image?"
• "Run cellpose on my TIFF and give me a cell count"
• "Segment my fluorescence image and export morphology metrics"

Output Structure

output_dir/
├── report.md
├── {stem}_measurements.csv
├── {stem}_cp_masks.tif
├── {stem}_seg.npy
├── figures/
│   ├── {stem}_cp_outlines.png
│   └── {stem}_histogram.png
└── reproducibility/
    ├── checksums.sha256
    ├── commands.sh
    └── environment.yml

Dependencies

• cellpose>=4.0 — cpsam model
• tifffile — TIFF I/O
• Pillow — PNG/JPG loading
• numpy — array ops
• matplotlib — figures
• scikit-image — regionprops metrics

Safety

• Local-first: no image data leaves the machine
• Every report includes the ClawBio medical disclaimer
• Reproducibility bundle (commands.sh, environment.yml, checksums.sha256) records the exact invocation, dependencies, and output integrity

Integration with Bio Orchestrator

Trigger conditions:

• Input is a TIFF/PNG/JPG microscopy image
• User mentions "cellpose", "segment", "cell counting", "microscopy"

Chaining partners:

• Future: export ROI centroids to spatial transcriptomics workflows

Citations

• Pachitariu, Rariden & Stringer (2025) Cellpose-SAM: superhuman generalization for cellular segmentation. bioRxiv 2025.04.28.651001 — CellposeSAM / cpsam model