Identification of Poisonous Fungi Basidiomycota Macro Based on Mobile Device Using Neural Network

Mushrooms are single or many-celled eukaryotic organisms without chlorophyll. Mushrooms are easy to grow in various places such as in the garden, in trees, in the forest and even in the yard. Mushrooms have different shapes and types. Mushrooms can be processed into a variety of delicious foods, even some mushrooms also have various health benefits. However, not all fungi can be consumed, there are some fungi that are poisonous and can cause death if consumed. The number of species and the lack of knowledge of Indonesian people about the macro Basidiomycota makes it difficult to identify fungi, so that many people are poisoned after consuming mushrooms. This result produced an application to detect the types of fungi found so that they could identify the types and benefits of mushrooms they found directly using an Android device. The mushroom identification process obtained from feature extraction was carried out on 1020 training data. Training data is obtained by authors from the internet. Feature extraction was carried out using two methods, namely 3D Color Vector Quantification for color extraction and Hu Moment for form extraction process. After that, the classification will be done using a neural network using Back propagation method. The result of this research is an application for detecting fungus types with accuracy 89.71%.