Fish Freshness Determination through Support Vector Machine

In this study, the fish freshness determination system used digital image processing to determine the freshness quality and shelf life span of the three most consumed fish in the Philippines namely: (1) milkfish (Chanos chanos), (2) round scad (Decapterus maruadsi) and (3) short mackerel scad (Rastrelliger brachysoma). Moreover, it used a method based on support vector machine (SVM) algorithm that would classify the redness of the fish’s eyes and gills as a measure of the fish freshness quality level. It will be able to determine the shelf life of a raw fish after it has been stored in a slurry ice. Standard images were set with technical assistance from the Philippines’ Bureau of Fisheries and Aquatic Resources (BFAR) that will be used as database of the program. The database for the network, which was successfully verified and approved by the aquaculturists from BFAR, includes 720 images for milkfish, 480 images for round scad, and 480 images for short mackerel scad.  The captured image of the fish to be tested will be processed by the MATLAB program.  It will be compared to the images in the database.  The results of the testing is compared with the manual sensory assessment done by the aquaculturists from BFAR achieving 98% accuracy in determining the freshness of the fish samples.