Grade Classification of Yellowfin Tuna Meat Using F-RCNN with Inception V2 Architecture

Yellowfin Tuna (Thunnus albacares), holds a significant economic importance as it is globally traded and widely consumed. Given its value, ensuring the quality of yellowfin tuna is vital. In this study, an Object Detection Convolutional Neural Network model F-RCNN with Inception V2 is developed to effectively classify tuna meat quality. Quality grades of the tuna were classified using letter grades: Grade A, Grade B and Grade C. It demonstrated an accuracy of 86.67% for Grade A, a perfect score of 100^% for Grade B, and 92.50% for Grade C. Overall, the model achieved an accuracy rate of 92.8% in accurately categorizing the different grades of tuna.

I. Introduction

Yellowfin tuna (Thunnus albacares), like all varieties of tuna, is a highly valued saltwater fish, renowned for its exceptional nutritional value and health benefits [1]. As a result of its nutritional composition, tuna is enjoyed globally in various forms, including fresh, frozen, and canned products [2]. Additionally, tuna holds a prominent position in the seafood market, being part of the high-value seafood groups such as cephalopods, shrimps, and lobsters [3]. The tuna industry plays a pivotal role in the economy, particularly in the Asia-Pacific region, where it employs over 6 million people, contributing significantly to regional economies and livelihoods [4]. Despite the slight decline in catches due to the COVID-19 pandemic, restrictions on fresh tuna exports and the sashimi market, the demand for tuna remains strong.