Detection of Mango Leaves Disease using CSUBW based Residual Attention with EfficientNet

Mangoes are one of the most popular fruits worldwide due to their taste and nutrition. Plant diseases may reduce plant quantity and quality. Disease management to improve crop quality and productivity requires early plant illness detection. Automation of plant disease detection reduces large farm surveillance. Since plants feed us, leaf problems must be diagnosed quickly and accurately. Our advanced learning method automatically detects leaf disease in several mango plant varieties. This study creates the RAM-EfficientNet, a novel illness classification network. The RAM block in EfficientNet extracts additional features and handles slight leaf lesion changes. Transfer learning was used to train EfficientNet on ImageNet to overcome the mango leaf disease sample size issue. The Cat Swarm Updated Black Widow Model (CSUBW), a hybrid optimization model, changes RAM-EfficientNet's weights to improve classification accuracy. Black Widow Optimization Algorithm (BWO) and Cat Swarm Optimization Algorithm (CSO) are combined to create the hybrid optimization model. The model's performance is then assessed to confirm its efficacy.