Colony Cat Management
A computer-vision project for community cat caretakers: identify individual cats from photographs, track colony members over time, and support humane TNR workflows.
What it solves
Community cat work depends on knowing which cats are already known, which are new, which are altered, and which may need follow-up. This project is designed as an assistant for caretakers, not an autonomous decision-maker.
Planned capabilities
Compare uploaded photos against known colony members using visual similarity.
Flag clipped or unclipped ear status to help track altered cats.
Store names, photos, markings, sightings, and health notes.
Use coat patterns, facial markings, ear shape, and tail features.
Support quick photo review and caretaker updates.
Capture low-confidence predictions for dataset improvement.
Build notes
- Start with labeled folders for clipped and unclipped examples.
- Add known-cat identity folders once enough photos exist per cat.
- Use embeddings first; avoid overfitting a tiny classifier.
- Keep confidence thresholds conservative.
- Design for caretaker review and corrections.