MANILA, Philippines – Crabifier, an application developed by a team from the De La Salle University (DLSU) aims to help crab farmers identify crabs which would yield the most profit for them when purchasing newly captured juveniles from traders.
The team explains the situation for crab farmers in the Philippines. Currently, there are 3 common mangrove crab species suitable for farming: Scylla serrata, Scylla tranquebarica, and Scylla olivacea.
Among the 3, the Scylla serrata, more commonly known as the giant mangrove crab, grows the fastest and the biggest, making them most appealing to the farmer.
However, when farmers buy newly caught crabs from traders, there is no easy way to tell the 3 species apart. As the team explains, “there is no obvious morphological marker to distinguish one species from the other.”
Dr. Chona Camille Abeledo, the woman behind the app, had been collecting alimango since 2012 as part of her dissertation. She told Rappler her original interest was in taxonomy, but the farmers eventually began "telling stories about how difficult it is to be sure that the juveniles they have will turn out to be their preferred species, and how sometimes traders would convince them they’re getting the species they need but it would turn out otherwise once the crabs have grown."
Basically, as there’s no big sign pasted on the back of the crab that says “Scylla serrata," farmers are buying blindly, hoping the crabs they purchase are of the serrata variety.
This is what Abeledo and her DLSU team seeks to address with their Crabifier app. The team determined that there were enough differences between the shapes of the frontal lobe spines of the 3 species. The frontal lobe spines are the spines found in the section between the crab’s two eyes. According to the team, the 3 species have different spine patterns – different enough that a smartphone can take a photo of a crab and identify what kind of crab it is among the 3.
The app makes use of a database of reference images to which new crab photos are compared. A farmer, for instance, takes a photo of a crab, and the app sends it to a server for classification. The app is assisted by artificial intelligence and neural networks – basically a smart database that continues to learn as it is fed new images – that should allow it to improve its speed and accuracy in identifying the species.
The team has also developed offline usability for the app. "We have created a compressed database within the app itself that it does not need internet connection to function; the database is within the phone instead of in a server," explains Abeledo.
With Abeledo, the team is currently comprised of developer Courtney Anne Ngo; student developer Marcus Ramos; field assistants Bienna Joaquin, Gerald Irigan and Karen Camille Perez; and Dr. Ma. Carmen Ablan-Lagman, the head of the practical genomics laboratory and adviser. Abeledo also credits developer Francheska Laguna, who has since left for further studies.
Know a promising startup, app idea, or novel implementations of technology? Email the author at email@example.com