Fish recognition model for fraud prevention using convolutional neural networks

Fraud, misidentification, and adulteration of food, whether unintentional or purposeful, are a worldwide growing concern. Aquaculture fisheries recognized as one the sectors most vulnerable to food fraud. Besides, series risks related health distrust between consumer popular market makes this sector develop an effective solution for fraud control. Species identification is essential aspect expose commercial Convolutional neural networks (CNNs) powerful tools image recognition classification tasks. Thus, objective study propose model fish species based on CNNs. After implementation comparison results CNNs, it was found that Xception architecture achieved better performance with 86% accuracy. It also possible build web application mockup. The proposal easily applied in other aquaculture areas such lobsters, shrimp, among seafood.