Performance Comparison of Convolutional Neural Network-based model using Gradient Descent Optimization algorithms for the Classification of Low Quality Underwater Images

Underwater imagery and analysis plays a major role in fisheries management and fisheries sciencehelping developing efficient and automated tools for cumbersome tasks such as fish species identification, stockassessment and abundance estimation. Majority of the existing tools for analysis still leverage conventionalstatistical algorithms and handcrafted image processing techniques which demand human interventions and areinefficient and prone to human errors. Computer vision based automated algorithms need a better generalisationcapability and should be made efficient to address the ambiguities present in the underwater scenarios, and can beachieved through learning based algorithms based on artificial neural networks. This paper research about utilisingthe Convolutional Neural Network (CNN) based models for under water image classification for fish speciesidentification. This paper also analyses and evaluates the performance of the proposed CNN models with differentoptimizers such as the Stochastic Gradient Descent (SGD),Adagrad, RMSprop, Adadelta, Adam and Nadam onclassifying ten classes of images from the Fish4Knowledge(F4K) database.