LASSR: Effective super-resolution method for plant disease diagnosis

The collection of high-resolution training data is crucial in building robust plant disease diagnosis systems, since such have a significant impact on diagnostic performance. However, they are very difficult to obtain and not always available practice. Deep learning-based techniques, particularly generative adversarial networks (GANs), can be applied generate high-quality super-resolution images, but these methods often produce unexpected artifacts that lower the In this paper, we propose novel artifact-suppression method specifically designed for diagnosing leaf disease, called Leaf Artifact-Suppression Super-Resolution (LASSR). Thanks its own artifact removal module detects suppresses considerable extent, LASSR much more pleasing, images compared state-of-the-art ESRGAN model. Experiments based five-class cucumber (including healthy) discrimination model show with generated by significantly boosts performance an unseen test dataset over 21% baseline, our approach than 2% better trained ESRGAN.