<![CDATA[The waste problem is a very complex global issue, especially in Indonesia. The volume of waste that continues to increase every year is a major challenge for the environment, health, and the economy. so it is necessary to conduct research related to smart waste management, namely, a concept of utilizing artificial intelligence in waste management by adopting image management techniques. Based on this, this study aims to compare the modeling of Covolutional Neural Network (CNN), GoogleNet, and Transfer Learning. The methods used in this study, CNN, GoogleNet, and Transfer Learning by utilizing data augmentation, activation functions, and transfer learning, are able to overcome the problem of limited data and reduce or avoid overfitting problems in modeling. The datasets used in this study are sourced from datasets built by the researcher himself and Kaggle datasets with a total of 300 samples consisting of 6 classes: Cardboard, Glass, Plastic, Metal, Paper, and Other/Trash. The results present that the transfer learning method is superior to other methods with accuracy, precision, recall, and f1-score, 100%. The contribution of this research is to enrich the literature in the field of machine learning and computer vision, develop more efficient models for limited datasets, and become a reference for future researchers who want to develop similar systems.]]>
Copyrights © 2026
