Enhancing Larval Classification Accuracy Through Hyperparameter Optimization in ResNet50 with Three Different Optimizers

Recent research has explored the application of deep learning models in the classification of Zophobas Morio and Tenebrio Molitor larvae, focusing on optimizing the choice of optimizer. This study aimed to investigate and compare the performance of the ResNet50 model using three different optimizers (Adam, Nadam, and Adagrad) for accurate classification of these larvae. We conducted an experimental study, training and evaluating the R model on a dataset of Zophobas Morio and Tenebrio Molitor larvae images using the mentioned optimizers. This study followed a quantitative approach to assess model performance. Our findings indicated that the Adam optimizer yielded the highest accuracy (approximately 97.1%), outperforming Nadam and Adagrad. However, recognizing Tenebrio Molitor still presented challenges, highlighting room for improvement. The results suggest that the choice of optimizer significantly influences the model's performance in larva classification. While Adam showed promising results, further enhancements are needed for accurate Tenebrio Molitor recognition. Future research should explore different deep learning architectures and larger datasets for improved classification outcomes.