Smart deep learning-based approach for non-destructive freshness diagnosis of common carp fish

Abstract

Assessment and intelligent monitoring of fish freshness are of the utmost importance in yield and trade of fishery products. Rapid and precise assessment of fish freshness using conventional methods considering the great volume of industrial production is challenging. In this study, instead of feature-engineering-based methods, a novel and accurate fish freshness detection is proposed based on the images obtained from common carp and by applying a deep convolutional neural network (CNN). To classify fish images based on freshness by the proposed approach, first, VGG-16 architecture was applied to extract features from fish images automatically. Then, a developed classifier block constructed by dropout and dense layers was utilized to classify fish images. The obtained results showed the classification accuracy of 98.21%, and in conclusion, the proposed CNN-based method has lower complexity with higher accuracy compared to traditional classification methods. This method is well-capable of monitoring and classifying fish freshness as a fast, low-cost, precise, non-destructive, real-time and automated technique.