An Advanced Rice Plant Disease Classification Through a Modified Efficient Net Deep Learning Model

A comprehensive analysis of Efficient Net model for rice plant health classification is presented in this research. Modern convolutional neural network architecture like this kind have become known for being efficient and scalable. Rice, being a staple food for a significant portion of the global population, faces numerous threats from pests and diseases, making early and accurate disease detection crucial for ensuring food security. We employed Efficient Net due to its adaptive scaling of depth, width, and resolution, which significantly enhances model performance on limited datasets. Our methodology involved implementation of a dataset of rice plant images labeled with various health conditions such as ‘healthy’ and ‘unhealthy.’ The Efficient Net model was used to classify these conditions accurately after sufficient training. Proposed model results are with other prevalent models in plant disease classification, such as VGG16, ResNet50, and InceptionV3, across multiple validation metrics, including accuracy, recall, precision and F1-score. The proposed system achieves Precision of 94.4%, Recall of 92.9%, and 95.11% of F1 score which outperformed the other existing models.