Few‐shot learning for plant disease recognition: A review

Abstract Monitoring plant diseases is essential for farmers to secure crop quantity and quality. Deep learning has recently been applied disease recognition help take prompt proper actions prevent reductions in Generally, deep requires a large‐scale dataset with supervised information annotated often by specialists. However, because collecting images natural environments difficult obtaining annotations from specialists costly, infeasible tasks. Few‐shot (FSL) an alternative using prior knowledge. Although FSL attracted considerable attention, comprehensive reports on the application of methods are required. Here, we introduce its applications recognition. We begin overview computer vision tasks machine FSL. provide practical examples applications. Utilizing these examples, describe different approaches data augmentation embedding, multitask learning, transfer meta‐learning. Further, summarize how models optimized performance reference existing studies. Finally, advantages disadvantages discussed, along potential challenges