FishSnap - Fishial Recognition

A mobile image recognition app that uses a custom fish category dataset as a pre-trained model for Core ML to classify what type of fish is in your photo. This allows smaller/poorly-funded fishers to quickly assess the nature/value of their catch and to track fish locations by the scans. This data collection will both facilitate aquatic marine studies and further improve the machine learning of the image recognition. Winner of the Fishackathon 2018 NYC this “fish identification” tool is designed to help minimize human errors, the need for expensive fish surveys on vessels, and allow smaller/poorly-funded fishers to better assess the nature/value of their catch. It is the hope that this app would be adopted by many regulatory and research agencies as well as the private sector to aid in the identification of fish species from citizen-science collected data. This application offers the flexible in implementation to allow for integration of other features in the near future.