<![CDATA[In this research, a classification of mushroom types based on digital image processing using a Convolutional Neural Network (CNN) is conducted. The method employs the EfficientNet-B4 architecture as the base model utilizing transfer learning and fine-tuning processes. The dataset consists of 3000 types of mushrooms, each categorized into 10 classes with 300 images per class. The CNN model is implemented using the Python programming language on Google Colab editor. Performance evaluation is carried out using accuracy, precision, recall, and F1-Score metrics to measure the model's performance. A comparison is made between all models with various training parameters, including identical and different settings. Additionally, the ratio of data splits, whether identical or different, is considered. Model 1, which utilizes a custom freeze layer and a data split ratio of 80% for training, 10% validation, and 10% testing, achieved the highest accuracy (90.00%), precision (90.09%), recall (89.63%), and F1-Score (89.59%) compared to other models. Therefore the implementation of a custom freeze layer to reduce the$ number of trainable parameters significantly impacts the accuracy level of the trained and tested model. Moreover, the determination of the data split ratio also slightly influences the accuracy level of the trained and tested model.]]>
Copyrights © 2025
