An Enhanced Approach to Recognize and Categorize Diseases Found in Mango Leaves Using Efficient-Net Architecture

Agriculture industry is a vast area of research. Although research in this field is a continuous process over decades still there is much more to explore. Detection of diseases on leaves of plants is an area to discover more. In this paper, variety of diseases on mango leaves have been discussed along with their causes and remedies. Further, we have detected whether the mango leaf is suffering from that particular disease or not. We have implemented our architecture using deep learning technique EfficientNet; a CNN based network and compared our results when implemented on other architectures such as ResNet, VGG, Xception and CNN. The results shown higher accuracy and lower error rates while detecting the mango leaf diseases as compared with other architectures.