Real-Time Recycle Waste Classification Using Deep Learning: A Convolutional Neural Network Model for Automated Waste Management

The advancement of segregation of recycle waste classification applications has been increased by the development of innovative solutions which utilise modern technologies for effective waste management and recycling practices. However, there have problems when most of the country are lack of garbage collection and mechanism facilities, traditional waste management is inefficient and costly, and most people are not aware of how to manage recycle waste correctly. Therefore, this paper presents and studies a smart recycle waste classification of Convolutional Neural Network (CNN) to automate the real-time recyclable waste classification for three types of waste such as metal, paper, plastic. Four phases that applied in this study which are data acquisition, pre-processing, processing (i.e CNN architecture model), and post-processing (i.e test and evaluation). A total of 3000 images have been experimented in this study. All image datasets have been captured in Malacca and collected from the Garbage Classification and Waste Segregation Image Dataset. The CNN architecture model has been applied in this study. To evaluate the quantitative predictive performance of the proposed model, confusion matrix, accuracy, precision, recall, and F1-score are computed with random-splitting and manual-splitting of train-test image dataset with the ratio of $70: 30,80: 20$, and $90: 10$. The results of this study showed the highest rate of $97.30 \%$: $93.67 \%$, $\mathbf{9 7 . 3 0 \%}$: 93.67%, $\mathbf{9 7 . 3 6 \%}$: 96.24%, and 97.26%: 96.18% for accuracy, precision, recall, and F1-score, respectively. Thus, it for is hoped to streamline waste sorting procedures and enable effective recycling techniques to deploy the smart recycle waste classification system, and significantly influence waste management initiatives, fostering sustainability and environmental protection. For the future work, this study may extend to experiment the other various type of recycle waste and experimented for various machine learning algorithm.