Enhanced Classification of Beech Bark Diseases Using CNN-MLP Fusion: A Multiclass Approach

Beech tree skin illnesses are common dangers to forest systems, needing the right and fast sorting methods to check how bad stages matter most for full handling plans. This research brings in a smart mix of methods using the powers of Convolutional Neural Networks (CNNs) and Multi-Layer Perceptron (MLP) to find out how bad beech bark diseases (BBD) are at different stages. Using a carefully picked set of 2000 very clear images, that show many different and hard-to-understand disease signs; this study looks deep into the careful sorting out of these levels of being serious. The model's great work is outstanding, with a really good overall rightness of 94.47% over five clear-cut levels of seriousness. Tough measuring tools, including accuracy, remembering and F1-score keep proving the model is good at exactly putting complicated differences in disease stages. The clear accuracy of the model shows it can pick up small details found in these steps. This points to how important it could be when dealing with tree care and also sickness control. These results show hopeful ideas for checking illnesses, finding them early, and knowing how best to help affected forest areas with the hard problems of BBD.