EfficientHACoTNet: A New Framework for Detecting Grape Leaf Diseases with Depthwise Separable Convolutions and Attention Mechanisms

Grapevine diseases are one of the major challenges in the viticulture industry, affecting not only the yield but also the quality of grapes. Timely and precise identification is essential for managing diseases and maintaining optimal crop health. This study presents EfficientHACoTNET, an innovative deep learning framework specifically developed for the efficient detection of diseases in grape leaves. The model utilizes depth-wise separable convolutions (DSC) to reduce computational complexity and attention mechanisms to enhance feature extraction from leaf images by striking a balance between high accuracy and computational efficiency. EfficientHACoTNET is built to work seamlessly on edge devices, making it a practical choice for real-time monitoring in precision agriculture. Extensive experiments conducted on specialized grape leaf disease datasets reveal that the proposed model outperforms existing state-of-the-art approaches in terms of both detection accuracy and inference speed. These findings underscore the potential of EfficientHACoTNET as a powerful tool for detecting major grapevine diseases, providing practical solutions for precision viticulture.