MarineMatch
2024, Senior, Interactive & Multimedia
Project Description: Overfishing is a huge problem across the world, and it has personally impacted my own community; this is what led MarineMatch to be born. Surprisingly, overfishing is not always intentional; new fishers often mistake endangered species for accepted catch, or perhaps are unable to recognize the species and decide to keep it. This motivated me to create MarineMatch, an app built in Android Studio in the Java language which uses image classification for users to identify 60+ different types of ocean animals and fish species. This was possible through a TensorFlow Lite deep-learning model with 100+ hours of training! The app classifies the image via an image recognition algorithm from the convolutional neural network, determining the species of the ocean creature and displaying it from the user – all within just split seconds. MarineMatch makes sure there is no excuse for irresponsible practices; fishers can easily identify the species of the fish and make responsible decisions! Through this project, 10 local fishers have already been able to contribute to sustainable fishing practices; being able to make informed decisions on whether to catch or release fish – maintaining the biodiversity of our waters. I really hope more people across North America can use MarineMatch, whether that be fishers, school students or even people passionate about ocean conservation. Additionally, I have linked an Instagram post for MarineMatch that also gives instructions on how to use the app: https://www.instagram.com/p/C3_TtqOJHw_/?img_index=4
