Stepping up Onion Classification Model using CNN Algorithm

Abstract

Traditional shallot classification methods, relying on visual inspection or conventional image processing, face limitations in dataset identification. To address the issues, we propose a CNN model for classifying shallot types. The study involves collecting a large dataset, preprocessing, and training the model with optimized parameters to maximize accuracy. By adjusting hyperparameters, we achieve a balance between accuracy and performance time. With 50 epochs and a batch size of 64, the model achieves over 80% accuracy in classification tests. These results demonstrate the effectiveness of CNN in shallot classification, outperforming traditional methods. Future work could explore advanced architectures like Generative Adversarial Networks (GAN) and Graph Convolutional Networks (GCN) to further enhance the model's performance.