Fish fraud detection
Fishazam
A tool that spots fish fraud by combining smartphone photos with spectrometer readings.
- Industry
- Food safety
- Discipline
- R&D and grants
- Our role
- Computer vision and ML
- Technology
- Python
- PyTorch
- XGBoost
- OpenCV
- ONNX
The challenge
Fish species are easy to swap at the counter and hard to tell apart by eye.
What we did
- Image analysis and data classification
- App showing the result with a confidence level and a data visualisation
