“One of the most complex hurdles we faced was how to maximise and condense a variety of information onto a tiny phone screen,” said Guan Chen Di, a computer science student at Nanyang Technological University, describing the development of SGFish, a new mobile application launched on Sept 17, 2026, by the Ichthyological Society of Singapore to help identify and track fish species in local waters.
Did You Know? Before the official launch of the SGFish app on the Apple App Store, the database was compiled using a Telegram bot dubbed Angler’s Pokedex, built by volunteer and cybersecurity professional Samuel Pua.
Society releases app to identify local marine life
The Ichthyological Society of Singapore, which counts more than 150 members, released the SGFish app on the Apple App Store to assist the public and anglers in identifying local marine life. Users can photograph their catch through the app, which runs an image recognition model capable of identifying more than 900 species while logging where each find was made. To ensure scientific accuracy, society volunteers and co-founders manually verify every submitted record before it enters the official database. Currently, the database holds about 2,300 observations covering over 400 species.
The mobile application was developed to address limitations found in the previous Telegram-based collection method, where time and location details could not be extracted from submitted images. Brian Sng, a software engineering student at Singapore Management University and lead developer of the app, built the computer vision model by collating and vetting more than 90,000 photos sourced from international databases, society members, and personal archives. It took two months to train the model, achieving an accuracy rate of 82.7%, which is more accurate than the 80% global fish identification accuracy rate of the US-based flora and fauna database iNaturalist.
App developers integrate alerts for protected species
The need for reliable identification tools became apparent in August at Bedok Jetty, where an angler mistook a globally endangered zonetail butterfly ray for an average stingray. Brian Sng identified the rare catch from a photograph and informed the angler of its value. Novice anglers frequently struggle to distinguish between dangerous and harmless fish or lack awareness regarding sustainable fishing practices. Consequently, developers integrated a pop-up alert feature that advises users on proper protocols when they scan protected, rare, or harmful species, such as releasing endangered catches back into the water.
To guard against incorrect classifications, the app displays the top five species predictions ranked by similarity, allowing the user to select the correct option. For field convenience in areas with poor cellular connectivity, the system includes a lightweight offline model alongside the primary online version. The application incorporates tide data to guide recreational anglers on optimal fishing times, alongside a comprehensive fish encyclopaedia designed to fit easily into a single phone view.
Citizen science apps supplement traditional diving surveys
Fish biologist Jeffrey Kwik noted that citizen science apps like SGFish provide vital data to supplement traditional diving and snorkelling surveys, which often struggle in Singapore’s murky waters. Spotting and counting fish is particularly difficult for well-camouflaged species or those residing in coral and seagrass habitats. Associate Professor Kwik, based at the Singapore Institute of Technology, explained that angler-contributed records help cover large survey areas efficiently, offering useful data for tracking fish distribution despite potential biases in hook sizes or bait types.
As the community continues to log observations, Andriel Cheong, co-founder of the Ichthyological Society of Singapore, expects the accumulated database to inform environmental impact assessments and assist authorities in coastal development planning. The developers also hope the general public will use the app to learn about seafood varieties found in local markets to encourage more conscious consumption. Future updates will bring the app to the Google Play Store and reintroduce gamification elements from the original Telegram bot, including a high-score board recognizing users who record the highest diversity of unique species.
Answers to common questions about the SGFish app
When was the SGFish app launched?
The SGFish app was launched on Sept 17, 2026, on the Apple App Store by the Ichthyological Society of Singapore.
How many species can the app identify?
The app uses an image recognition model capable of identifying more than 900 species of fish.
What accuracy rate does the app’s computer vision model achieve?
Trained over two months using a dataset of more than 90,000 photos, the model achieves an accuracy rate of 82.7%.
What species of marine life have you encountered unexpectedly in local waters?
Keep reading