ABSTRACT With the rapid rise of industrialization and modernization, improper waste disposal has emerged as a significant environmental challenge. Addressing this issue necessitates the development of an automated system for sorting and recycling waste to promote sustainable waste management practices. Recent progress in deep learning, particularly in image classification, offers promising tools for such applications. In this research, we introduce RWCNet (Recyclable Waste Classification Network)—a novel deep learning framework designed to categorize waste into six distinct types using the TrashNet dataset, which comprises 2,527 labeled waste images. The proposed model is rigorously evaluated through both quantitative and qualitative analyses and benchmarked against several cutting-edge waste classification models. RWCNet achieves an impressive overall accuracy of 95.01%, surpassing multiple state-of-the-art techniques. Moreover, it attains high F1-scores across all waste categories: 97.24% for cardboard, 96.18% for glass, 94% for metal, 95.73% for paper, 93.67% for plastic, and 88.55% for litter. To further validate its performance, Score-CAM-based saliency maps are employed, offering visual interpretation of the model's focus areas during classification. The findings affirm the robustness and precision of RWCNet, positioning it as a highly effective solution for automated waste categorization and recycling systems. Keywords: waste classification, deep learning, RWCNet, TrashNet dataset, image classification, recyclable waste, Score-CAM, saliency maps, automated waste management.
Pujar et al. (Sun,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: