Singapore’s Ichthyological Society has launched a new mobile application designed to help users identify and log local fish species, aiming to enhance public awareness of the region’s marine biodiversity. The app, called SGFish, was made available on the Apple App Store on September 17 and features an image recognition model capable of identifying over 900 fish species found in Singapore’s waters.
The initiative was partly inspired by a recent incident at Bedok Jetty, where an angler misidentified a globally endangered zonetail butterfly ray as a more common stingray species. The discovery, brought to light by a local fishing enthusiast and software engineering student, highlighted the need for a reliable tool to support accurate fish identification, particularly for rare or protected species.
Users of SGFish can take photos of their catches through the app, which then generates a list of the top five possible matches ranked by similarity. To ensure data accuracy and scientific integrity, all submitted records are manually verified by volunteers from the Ichthyological Society of Singapore (ISS) before being added to the database. The app also includes notifications that advise anglers on appropriate actions if they catch protected or harmful species, such as releasing endangered fish back into the wild.
The database supporting the app contains roughly 2,300 verified observations covering more than 400 species. This data was initially collected through a Telegram bot called Angler’s Pokedex, developed by cybersecurity professional and volunteer Samuel Pua. However, limitations with the bot—such as the inability to extract time and location information from images—prompted the team to develop the more advanced app.
Lead developer Brian Swng, a student at Singapore Management University, created the app’s computer vision model by compiling and vetting over 90,000 fish photographs from international sources, society members, and personal collections. After two months of training, the model achieved an accuracy rate of 82.7 percent, surpassing that of the U.S.-based iNaturalist platform, which has about 80 percent accuracy for fish worldwide.
Experts have noted that initiatives like SGFish can supplement traditional fish population surveys, which are often challenged by Singapore’s murky waters and the camouflage abilities of certain species. Jeffrey Kwik, a fish biologist and associate professor at the Singapore Institute of Technology, emphasized the value of data collected from anglers in covering large survey areas and tracking fish distribution, despite potential biases linked to fishing methods.
ISS co-founder Andriel Cheong expressed hope that the crowd-sourced database will prove useful in environmental impact assessments and coastal development planning by providing authorities with better insights into local marine ecosystems. Beyond recreational fishers and divers, the team also encourages the general public to use the app to learn about local fish species, including those sold in markets, to foster more informed and conscious seafood consumption.
Future updates to SGFish aim to expand accessibility to Android users via the Google Play Store and to integrate gamification features from the original Telegram bot, such as leaderboards recognizing users who photograph the highest diversity of species.