Machine Learning

LuniFish: Machine Learning Applied to Fish Species Recognition

Factordev contributed to the development of IASAI (Image Analytics for Seafood machinery using AI), focusing on the integration of Artificial Intelligence and Machine Learning with the aim of making fish species recognition more efficient.
From this experience, LuniFish, was born – an advanced system that brings innovation to the seafood industry

Machine Learning e Computer Vision

Thanks to an advanced Machine Learning and Computer Vision system, LuniFish is able to recognize fish species in real-time and estimate their size by analyzing images captured in seafood markets

Automation and precision : transforming the seafood industry

LuniFish uses unsupervised Deep Learning technologies for size estimation and supervised Machine Learning models for species recognition

This approach allows to

How LuniFish works:

  • Fish images are captured through IPCAMs installed on industrial machinery.
  • The system analyzes the data and automatically classifies species and size.
  • The information is processed in the Cloud , continuously improving the model through Reinforcement Learning.
  • The system has demonstrated a reduction in execution times and optimization of processing and packaging operations.
Revolutionize your seafood company's process management with LuniFish machine learning

LuniFish: the future of seafood management is already a reality

By adopting LuniFish, companies in the seafood industry can automate, speed up, and improve their operations, bringing concrete and sustainable innovation to the market. Moreover, thanks to its versatility, the system can be adapted to other sectors of the agri-food industry, supporting the automation of classification and traceability processes for various types of products

Want to know how LuniFish can improve your business?


The project is hosted on Zenodo, a platform dedicated to the publication of co-funded projects.
Thanks to Zenodo, it can receive a DOI (Digital Object Identifier), making it easier to cite in other research and publications