The Art of Multi-Classification: Detecting Rice Sheath Rot Disease Severity Levels using a Hybrid CNN-SVM Model

Rice sheath rot is a dangerous disease that affects rice crops all over the globe. It is responsible for substantial yield losses and considerable economic damage. To effectively handle an illness, early discovery and accurate classification of the disease’s severity are both necessary. For this research, the authors suggested a method that combines Convolutional Neural Networks (CNN) and Support Vector Machines (SVM) for the diagnosis of the illness known as rice sheath rot. The suggested method consists of two stages: a binary classification step to determine whether or not the illness is present, and a multi-classification stage to categorize the severity to which the disease has manifested itself into five distinct stages. The current study evaluates the performance of the suggested strategy on a dataset consisting of images of rice sheath rot disease and compares it to other approaches that are considered to be state-of-the-art for disease identification in rice crops. When it came to detecting and categorizing the illness in rice crops, the suggested method was able to achieve an accuracy of 95.2% overall. According to the findings of the research, the suggested hybrid strategy consisting of CNN and SVM will be more successful in identifying rice sheath rot disease than other methods that are considered to be state-of-the-art. Further research is needed to validate and extend the proposed approach to other plant diseases and to ensure its generalizability to other variations of rice sheath rot disease.