PlantViQ: Disease Recognition Across Varied Environments with Vision Transformer and Quadrangle Attention

Accurate plant disease detection in real-world agricultural settings poses a significant challenge due to the complex and variable nature of field environments. To address this challenge, we introduce PlantViQ. This deep learning model leverages the strengths of both convolutional neural networks (CNNs) for local feature extraction and Transformers for capturing global context. This enables PlantViQ to overcome the limitations of traditional CNN-based approaches, particularly in scenarios with small-sample datasets and challenging environmental conditions. As demonstrated by its exceptional F1-score of 88.52% on cassava leaf disease and 97.21% on rice leaf disease, PlantViQ has the potential to revolutionize crop management practices by enabling early and accurate disease detection even in the most challenging real-world environments.