Classification and Identification of Potato Leaf Disease Using Ensembling of Attention-Aided Deep Neural Networks Through Feature Fusion

In terms of production, the potato is the most grown crop after wheat and rice in India. Foliar diseases, most notably Late Blight and Early Blight, expose the yields to risk. These diseases, if left undiagnosed, can pose grave economic losses along with a severe drop in productivity. Early and accurate detection, therefore, is important. This paper presents a deep learning-based technique that uses the fusion of features from multiple Convolutional Neural Network (CNN) models enhanced with attention mechanisms for refined disease classification. By distinguishing healthy leaves and leaves suffering from Early and Late Blight, the system provides timely advice to farmers. The proposed method, which outperforms others, achieves a remarkable 97 % classification accuracy.