CRABIFIER is based on a problem-solving system that uses image analysis and genetic marker technology to look for discrete visual differences in organisms that are hard to identify. The system is transformed into a mobile application using convolutional neural networks that can quickly do the identification on the field at the touch of a button. It is currently used for juvenile mangrove crabs but can eventually be used for any organism at any developmental stage, dependent on the needs of the user.
CRABIFIER is the first iteration of this problem-solving system. It is a mobile app used for species identification in mangrove crabs.
The mangrove crab industry requires efficient allocation of resources. The most common practice involved is the fattening of captured juveniles from the wild or purchase from traders who source it from other locations. Productivity of farms is limited by the size of the available pond, the cost of feeds and maintenance of the farm. Since fishers cannot verify if the juveniles they have are of the species they need, they overstock theirs ponds to compensate for expected loss. Traders sometimes claim that they have their own traditional means of species identification, but fishers have asserted that they would like to have a more accurate means to know for themselves.
Of the three mangrove crab species in the Philippines, Scylla serrata, or the giant mangrove crab, is capable of growing faster and bigger compared to S. tranquebarica and S. olivacea. There is no obvious morphological marker to distinguish one species from the other. Genetic markers were used to identify to species juvenile mangrove crabs and image analysis was used to find consistent physical markers on the shell of the crab to represent the species. This process won the 2015 Outstanding Research and Development award of the Philippine Council for Industry, Energy and Emerging Technology Research and Development.
The problem-solving system found subtle differences in the crown of the mangrove crab shell useful for species differentiation. The system made use of genetic barcodes unique to the species of individual crabs, and image analysis that converts geometric shapes in the crown into Fourier components (FC). Multivariate analysis on the FC was then done to check for species clusters and reference images were traced from the found groups.
Dr. Ma. Carmen Ablan-Lagman and Dr. Chona Camille VinceCruz-Abeledo receiving the 2015 Outstanding Research and Development Award for the species identification system in mangrove crabs.
A juvenile crablet being measured prior to processing.
To make the system mobile and useful in the field, the identified reference images were processed using convolutional neural networks and transformed into a mobile application called CRABIFIER. Since then, CRABIFIER has been field-tested in fish ponds and fishermen communities in Binangonan, Zambales and in Buguey, Cagayan, where the first training of fishermen on the use of the app was held.
A giant mangrove crab welcomes visitors to Buguey, Cagayan.
Ms. Courtney Anne Ngo showing fishermen in Buguey, Cagayan how to use CRABIFIER.
Members of the Practical Genomics Laboratory (PGL) with representatives of the Local Government of Buguey, Cagayan during a field testing of Crabifier. From Left to Right: Ms. Courtney Anne Ngo, developer of CRABIFIER, Fragel Mendoza, mangrove crab farm owner, Gerald Irigan and Biena Joaquin, Research Associates of PGL, Danilo Rumpon, Municipal Agriculturist, Mayor Lloyd Antiporda, and Jun Mendoza, another mangrove crab farm owner.