Deep Neural Networks for Accurate Identification of Edible and Poisonous Fungi

To keep ourselves from harm and safeguard of our health it is most important to distinguish between harmful and edible fungi. This paper presents a thorough study on classifying edible fungi as against inedible ones using deep learning. In particular a Deep Convolutional Neural Network (DCNN) model has 7 layers, and this task is carried out by it because it is good at extracting the subtle features from fungus images. The dataset used in this text is a diverse array of photos, each representing a different kind of edible or poisonous fungi. Round the clock, every effort has been made to label the image. This is to guide further learning while the CNN model is rigorously trained on this dataset. Every technique available to enhance generalization and help prevent over-fitting is deployed, such as using data augmentation methods. At the end of rigorous training and validation, this CNN model duly racked up an impressive accuracy rate: 83%. This high level of accuracy demonstrates the effectiveness of carefully classified fungal specimens using deep learning methods, and as a result the food security worker can make rational choices when consuming nutritious fungi. The methods and results presented in this paper will help greatly to advance the classification of fungi as either edible or harmful; furthermore, it provides a reliable automatically operated system do he same. Such advances are likely to greatly ease the load on food safety inspectors and give important information to foragers in the wild who come across wild mushrooms, as well as those of us in cooking-know it all who actually know little about what type of fungi we are dealing with.