A Novel MobileInceptionNet Architecture for Apple Leaf Disease Classification

Diagnosis of agricultural plant leaf diseases is a significant, but a time-taking process, if performed manually. Therefore, automated systems are much required for the early diagnosis of the leaf diseases. Recent advancements in Deep Learning(DL) have accelerated the development of these autonomous systems. This study introduces a novel deep learning model designed, particularly to classify apple leaf diseases accurately, hence, addressing the critical aspect of agricultural management. The proposed model combines the efficiency and lightweight-nature of the MobileNet architecture with the strong feature extraction properties of Inception module and hence, is named MobileInceptionNet. To assess the efficiency of the model, metrics such as recall, precision, and f1-score are used. An average precision, recall and F1-score of 0.947, 0.946 and 0.946 respectively were recorded for experimental results across thirteen different classes. With a test accuracy of 94.99%, the proposed model has demonstrated its potential in improving automatic diagnosis systems for apple leaf diseases.