International Journal of Advanced Research in Computer and Communication Engineering

Abstract: This paper presents a Convolutional neural network-based automated waste segregation system. For efficient waste management. Food waste, metal, plastic, and paper are the four categories into which the model divides waste images. It uses a deep learning technique to classify images. Training and testing are conducted using the Waste Segregation Large Dataset from Kaggle, which includes more than 56,000 labeled images. To efficiently extract and classify features, the CNN architecture includes multiple convolutional, pooling, and dense layers. The suggested system's consistent accuracy demonstrates CNNs' high level of precision in waste segregation automation. Additionally, a variety of data improvement techniques are applied to lessen overfitting and boost the model's generalization ability. The model's strength is ensured by assessing its performance using metrics such as accuracy, precision, recall, and F1-score. The system can integrate into smart waste management setups, where cameras and sensors automatically capture and classify waste in real time. This automated process reduces manual work and human error while optimizing recycling tasks by ensuring precise sorting. Overall, the proposed approach helps promote sustainable recycling, conserve resources, and support cleaner urban areas.

Keywords: Waste Segregation · Convolutional Neural Network · Image Classification · Deep Learning · Smart Waste Management

Downloads: | DOI: 10.17148/IJARCCE.2025.141146