A Comparative Study of Deep Learning Models for Guava Leaf Disease Detection

Plant diseases are regarded as one of the major variables affecting food production and reducing production losses. Early identification and disease diagnosis of fruit crops are essential for the food industry to increase production. The development of the healthy fruit sector is highly dependent on plant diseases. Guava fruit is a good source of fiber, calcium, pectins, and vitamin C. Guava disease has become a major issue for guava production, which has a negative impact on farmers' socioeconomic growth. Infections that harm guava plants and fruits pose the greatest obstacles, although they can be avoided by early identification. Deep learning has recently been widely used to automatically identify a variety of diseases that cannot be seen with the human eye. In this study, convolution neural models are suggested to use leaf images to detect guava disease. The presented models can be used to categorize the main guava leaf diseases, including Canker, Dot, Healthy, Mummification and Rust. The kaggle dataset is used to gather images of guava with the disease Canker, Dot, Healthy, Mummification and Rust. In our work, CNN models like DenseNet121, DenseNet169, InceptionV3, and Xception are used to categorize the diseases that affect guava leaves. The experimental findings indicate that, with an accuracy of 96.12%, the DenseNet169 model outperformed other models. When the model's effectiveness is assessed using performance metrics like precision, recall, and F1 score, the results demonstrate that DenseNet169 gives excellent results.