ML based Fish Species Image Classification

Fish species classification is a critical task in marine biology and fisheries management, providing essential data for ecological studies and conservation efforts. This paper explores the application of deep learning techniques to classify fish species using image data. We employed four pre-trained convolutional neural network (CNN) models—VGG19, InceptionV3 and MobileNetV2 to categorize images into ten distinct fish species: Black Sea Sprat, Gilt-Head Bream, Horse Mackerel, Red Mullet, Red Sea Bream, Sea Bass, Shrimp, Striped Red Mullet, Trout, and Fighting. The dataset, sourced from Kaggle, comprises 7,000 color images, equally distributed across the ten species, and was split into $70 \%$ for training and $30 \%$ for validation. Each image, ranging in size from 100 KB to 200 KB and in PNG format, was used to train and validate the models. Our results indicate that the chosen deep learning models can achieve high accuracy in classifying fish species, with InceptionV3 and MobileNetV2 showing particularly strong performance. This study demonstrates the potential of deep learning in automating fish species identification, which can significantly aid in biodiversity monitoring and resource management.