An Enhanced Precise and Severity Evaluation of Multi Class Plant Disease Prediction Using Epistemic Neural Network with Spike-Driven Transformer

Abstract

One of the primary issues in the agricultural sector is Multiclass Plant Diseases (MPDs). For automatically identifying them is essential for keeping an eye on the plants. Plants' leaves display the majority of disease signs, however leaf diagnosis by specialists in labs is expensive and time-consuming. Numerous techniques are used for predicting Multiclass Plant Diseases. However, there are numerous drawbacks to the existing approaches, including a high error rate and low accuracy. To overcome the before mentioned  problem,  Epistemic Neural Network with Spike-Driven Transformer using Atomic Orbital Search Algorithm (EcNN-SeDTr-AOSA) is proposed for predict the multiclass plant diseases with high accuracy. In this input data is taken from Groundnut dataset. To reduce noise in the input data, Iterative Self-Guided Image Filtering (ISGIF) is proposed. Following that, the pre-processed images undergo feature extraction using Modified ResNet-152 (M-ReNt-152). After that classification is done by using Epistemic Neural Network with Spike-Driven Transformer (EcNN-SeDTr) and optimized using Atomic Orbital Search Algorithm (AOSA) for forecasting the multiclass plant diseases with outstanding accuracy and effectiveness. The efficiency of the proposed EcNN-SeDTr-AOSA is analyzed using Groundnut dataset and attains 99.5% accuracy, 99.3 % recall and attains better results in comparison with the existing techniques. The outcomes of the proposed technique showed that it could improve the crop yield, reduced losses and early disease detection of the Multi Class Plant Disease Prediction method.