Convolutional Neural Network for the Detection of Mushroom Disease Diagnosis

Mushrooms are commonly consumed and offer various nutrients, antioxidants, proteins, minerals, and vitamins, holding medicinal value. The fleshy fruiting bodies of fungi, grown on the soil, are known as mushrooms. Unlike other crops, mushrooms are susceptible to microbial, bacterial, and viral infections, with common diseases such as wet bubble, dry bubble, cobweb, and bacterial blotches. This paper aims to develop a unique approach using machine learning models for detecting and classifying mushroom diseases, reducing human involvement, and comparing model performance. Initial image pre-processing is conducted, followed by texture feature extraction using GLCM. Various classifiers, such as Multiclass Support Vector Machine (MSVM) and Random Forest, are applied to categorize mushroom diseases. Experimental results reveal that the Random Forest algorithm outperforms MSVM, achieving an accuracy rate of 82% compared to 76%. Performance evaluation, using a confusion matrix, includes parameters like precision, recall, and F1 score.