Image super resolution using Enhanced Attention Block (EAB) for Grape Leaf Disease Detection

Accurate disease diagnosis in grape leaf analysis is heavily reliant on high-resolution (HR) images. To meet this need, we propose a model that incorporates the Enhanced Attention Block (EAB) to generate HR images for improved disease detection. The EAB employs advanced attention mechanisms and residual learning to effectively capture critical, spatial and channel-wise relationships. Proposed model is assessed using PSNR (Peak Signal-to-Noise Ratio) and SSIM (Structural Similarity Index) metrics, yielding PSNR scores of 31. 902dB, 32.785dB, and 34. 853dB, and SSIM score of 0.8709, 0.9058, and 97.46 for factors 2,4 and 6 respectively. This proposed approach holds great potential for widespread implementation in disease detection across various plant species, allowing for timely and accurate interventions for plant health management and increasing industry productivity.