A Deep Learning-Based Fine-Tuned ResNet50 Model for Multiclass Mango Leaf Disease Classification

The cultivation of mango in tropical and subtropical climates faces significant challenges due to diseases that can impact crop productivity and quality. The ResNet50 model is a convolutional neural network (CNN)-based model which has 50 layers. This model has emerged as a potential tool for identifying mango leaf diseases. In various agricultural problems, it has shown outs tanding performance by performing excellent image categorization tasks. It is also suitable for identifying various types of diseases affecting mango leaves, such as anthracnose, powdery mildew, and bacterial black spots. In the proposed work, the potential of the ResNet50 model to transform disease control practices within the agricultural sector is highlighted It can identify diseases early and reduce the need for human inspection, by providing valuable data insights to enhance disease management processes. Implementing the ResNet50 model for early dis eas e detection in mango leaves is essential. It is advantageous to use the ResNet50 model for mango farming, scientific investigation, and agricultural expertise as half of the world takes mango in their basic diet. It facilitates prompt intervention and enhanced disease detection, fostering augmented mango harvests and promoting the long-term viability of the agriculture sector. The proposed Res Net50 model achieved a remarkable classification accuracy of 99.7% after fine-tuning on the Kaggle-based dataset of mango leaf images. Through rigorous training and validation procedures, the model demonstrated outstanding performance in discerning various types of mango leaf diseases, such as anthracnose, powdery mildew, and bacterial black spots. The model's exceptional precision not only demonstrates its efficacy but also presents a viable ins trument for early dis eas e detection, intervention, and ultimately, the conservation of mango crop producti vity.