<![CDATA[Waste, encompassing both domestic and industrial materials, presents a significant environmental challenge. Effectively managing waste requires accurate identification and classification. Convolutional Neural Networks (CNNs), particularly the Residual Network (ResNet) architecture, have shown promise in image classification tasks. This research aims to utilize ResNet to identify types of waste, contributing to more efficient waste management practices. The ResNet101 architecture, comprising 101 layers, is employed in this study for waste classification. The dataset consists of 2527 images categorized into six classes: Cardboard, Glass, Metal, Paper, Plastic, and Trash. The ResNet model is pre-trained, leveraging existing knowledge to enhance classification accuracy. The dataset is divided into training and testing sets to evaluate the model's performance. The testing results, evaluated using a Confusion Matrix, demonstrate strong performance in waste classification. The ResNet101 model achieves 92% accuracy in detecting inorganic waste objects within the training dataset and maintains a high accuracy of 90% on the testing dataset. This indicates the effectiveness of the ResNet architecture in accurately identifying various types of waste, contributing to improved waste management efforts. he utilization of ResNet101 for waste classification yields promising results, with high accuracy rates observed across both training and testing datasets. By effectively identifying types of waste, this approach facilitates more efficient waste management practices, enabling better resource allocation and environmental conservation. Further research and application of CNN architectures in waste management could lead to enhanced sustainability efforts and improved waste-handling strategies.]]>
Copyrights © 2024
