Smart Waste Classification Using Deep Learning for Sustainable Waste Management | IJCT Volume 13 – Issue 3 | IJCT-V13I3P125
International Journal of Computer Techniques
ISSN 2394-2231
Author
Simran Ghatore, Amritpreet Kaur, Vijay Chhabra
Abstract
Effective waste classification is a critical cornerstone of modern sustainable waste management systems, particularly in the context of rapidly urbanizing smart cities. The exponential increase in solid waste generation worldwide, automated solutions capable of categorizing waste with high accuracy and low computational overhead. This paper presents a comprehensive comparative study and hybrid deep learning framework for automated waste classification using the TrashNet benchmark dataset, which comprises 2,527 labeled images distributed across six waste categories: cardboard, glass, metal, paper, plastic, and trash. Three state-of-the-art convolutional neural network architectures—MobileNetV2 [1], EfficientNet-B0 [2], and ResNet-18 [3]—are systematically evaluated through transfer learning on the TrashNet corpus [4]. Following rigorous individual performance assessment, the two highest-performing base models are hybridized into a novel ensemble architecture that leverages feature-level fusion and probability averaging to achieve superior classification accuracy. Experimental results demonstrate that the proposed hybrid model attains an accuracy of 94.7%, outperforming all individual baseline architectures by a statistically significant margin. The proposed framework exhibits strong generalization capability, robustness to class imbalance, and suitability for deployment in resource-constrained edge computing environments encountered in smart city waste infrastructure. The study further contributes detailed confusion matrix evaluations, and per-class precision-recall assessments. The findings confirm that intelligent deep learning-based waste sorting systems can substantially advance the Sustainable Development Goal (SDG) 11 and SDG 12 targets relating to sustainable cities and responsible consumption [5].
Keywords
Deep Learning; Hybrid Ensemble; Smart Cities; Sustainable Waste Management; TrashNet, Transfer Learning
Conclusion
This paper presented a comprehensive investigation of deep learning-based waste classification using the TrashNet benchmark dataset. Three state-of-the-art CNN architectures—MobileNetV2, EfficientNet-B0, and ResNet-18—were systematically evaluated under a unified transfer learning protocol, achieving test accuracies of 91.3%, 92.6%, and 89.7%, respectively. Building upon these individual results, a novel hybrid model was proposed that combines the two best-performing architectures (EfficientNet-B0 and MobileNetV2) via weighted probability averaging, achieving a superior test accuracy of 94.7%—the highest reported result on TrashNet.
Ablation experiments confirmed the individual contributions of data augmentation, class-weighted loss, and end-to-end fine-tuning, providing actionable guidelines for practitioners working with small waste image datasets. Per-class analysis revealed that the paper and metal categories are most accurately classified, while the trash and plastic categories present the greatest challenges due to class imbalance and inter-class visual similarity, respectively.
The proposed hybrid model exhibits strong suitability for smart city edge deployment, with an inference latency of 18.1 ms on commodity embedded hardware and a 34 MB model footprint. These characteristics, combined with state-of-the-art accuracy, position the proposed framework as a practical and effective solution for automated waste sorting systems in urban environments. By enabling more accurate and scalable waste classification, this work contributes to the global pursuit of sustainable waste management and the realization of SDG 11 and SDG 12 targets.
References
[1] M. Sandler, A. Howard, M. Zhu, A. Zhmoginov, and L.-C. Chen, “MobileNetV2: Inverted residuals and linear bottlenecks,” in Proc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit. (CVPR), Salt Lake City, UT, USA, 2018, pp. 4510–4520.
[2] M. Tan and Q. V. Le, “EfficientNet: Rethinking model scaling for convolutional neural networks,” in Proc. 36th Int. Conf. Mach. Learn. (ICML), Long Beach, CA, USA, 2019, vol. 97, pp. 6105–6114.
[3] K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” in Proc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit. (CVPR), Las Vegas, NV, USA, 2016, pp. 770–778.
[4] G. Thung and M. Yang, “TrashNet: Dataset of garbage images,” GitHub repository, 2016. [Online]. Available: https://github.com/garythung/trashnet.
[5] United Nations, “Transforming our world: The 2030 agenda for sustainable development,” Resolution A/RES/70/1, United Nations General Assembly, New York, NY, USA, 2015.
[6] S. Kaza, L. Yao, P. Bhada-Tata, and F. Van Woerden, “What a waste 2.0: A global snapshot of solid waste management to 2050,” World Bank Group, Washington, DC, USA, 2018.
[7] A. Zanella, N. Bui, A. Castellani, L. Vangelista, and M. Zorzi, “Internet of things for smart cities,” IEEE Internet Things J., vol. 1, no. 1, pp. 22–32, Feb. 2014.
[8] I. Goodfellow, Y. Bengio, and A. Courville, Deep Learning. Cambridge, MA, USA: MIT Press, 2016.
[9] M. Abdallah, M. Abu Talib, S. Feroz, Q. Nasir, H. Abdalla, and B. Mahfood, “Artificial intelligence applications in solid waste management: A systematic research review,” Waste Manag., vol. 109, pp. 231–246, May 2020.
[10] S. J. Pan and Q. Yang, “A survey on transfer learning,” IEEE Trans. Knowl. Data Eng., vol. 22, no. 10, pp. 1345–1359, Oct. 2010.
[11] M. A. Ganaie, M. Hu, A. K. Malik, M. Tanveer, and P. N. Suganthan, “Ensemble deep learning: A review,” Eng. Appl. Artif. Intell., vol. 115, p. 105151, Oct. 2022.
[12] A. Krizhevsky, I. Sutskever, and G. E. Hinton, “ImageNet classification with deep convolutional neural networks,” Commun. ACM, vol. 60, no. 6, pp. 84–90, Jun. 2017.
[13] K. Simonyan and A. Zisserman, “Very deep convolutional networks for large-scale image recognition,” in Proc. 3rd Int. Conf. Learn. Represent. (ICLR), San Diego, CA, USA, 2015.
[14] C. Szegedy et al., “Going deeper with convolutions,” in Proc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit. (CVPR), Boston, MA, USA, 2015, pp. 1–9.
[15] M. Yang and G. Thung, “Classification of trash for recyclability status,” CS229 Project Report, Stanford University, Stanford, CA, USA, 2016.
[16] O. Awe, R. Menon, A. Pakala, and P. Olaosebikan, “Smart waste categorization and recycling using deep learning convolutional neural networks,” in Proc. Int. Conf. Comput. Sci. Appl. (ICSA), 2017.
[17] A. V. Vo, T. V. Nguyen, and T. T. Pham, “Waste classification using transfer learning of deep convolutional neural networks,” in Proc. Int. Conf. Green Technol. Sustainable Dev. (GTSD), Ho Chi Minh City, Vietnam, 2020, pp. 363–368.
[18] W. Mao, J. Wang, and Z. Xue, “An ELM-based model with sparse-weighting strategy for sequential data imbalance problem,” Int. J. Mach. Learn. Cybern., vol. 8, no. 4, pp. 1333–1345, Aug. 2017.
[19] Z. Kang, C. Yang, Z. Yin, and H. Zhao, “Multi-scale feature fusion for waste classification,” IEEE Access, vol. 10, pp. 28748–28757, 2022.
[20] C. Bircanoglu, M. Atay, F. Beser, O. Genc, and M. A. Kizrak, “RecycleNet: Intelligent waste sorting using deep neural networks,” in Proc. IEEE Innov. Intell. Syst. Appl. Conf. (ASYU), 2018, pp. 1–7.
[21] S. Serafini, J. Park, and M. Brathwaite, “Sustainable smart city waste management using lightweight mobile deep learning,” in Proc. IEEE Int. Smart Cities Conf. (ISC2), 2020, pp. 1–6.
[22] L. Rokach, “Ensemble-based classifiers,” Artif. Intell. Rev., vol. 33, no. 1, pp. 1–39, Jan. 2010.
[23] Z.-H. Zhou, Ensemble Methods: Foundations and Algorithms. Boca Raton, FL, USA: CRC Press, 2012.
[24] M. S. Rahman, F. Islam, A. Hossain, and M. T. Islam, “An ensemble approach for waste classification using deep neural networks,” IEEE Access, vol. 9, pp. 128858–128869, 2021.
[25] Y. Shi, X. Zhang, R. Wang, and L. Yin, “Feature-level fusion of EfficientNet and ResNet variants for fine-grained waste classification,” Comput. Electr. Eng., vol. 103, p. 108301, Oct. 2022.
[26] T. Rashid, S. Nisar, and F. Al-Turjman, “A secure and intelligent framework for waste management in smart cities,” IEEE Internet Things J., vol. 9, no. 7, pp. 4887–4898, Apr. 2022.
[27] C. Bфункci, E. Zanella, and M. Moreno, “An IoT-based framework for smart waste monitoring in urban environments,” Sensors, vol. 21, no. 6, p. 2012, Mar. 2021.
[28] P. Domingos, “MetaCost: A general method for making classifiers cost-sensitive,” in Proc. 5th ACM SIGKDD Int. Conf. Knowl. Discov. Data Min. (KDD), San Diego, CA, USA, 1999, pp. 155–164.
[29] A. Krizhevsky, “One weird trick for parallelizing convolutional neural networks,” arXiv preprint arXiv:1404.5997, 2014.
[30] J. Hu, L. Shen, and G. Sun, “Squeeze-and-excitation networks,” in Proc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit. (CVPR), Salt Lake City, UT, USA, 2018, pp. 7132–7141.
[31] X. Glorot and Y. Bengio, “Understanding the difficulty of training deep feedforward neural networks,” in Proc. 13th Int. Conf. Artif. Intell. Stat. (AISTATS), 2010, pp. 249–256.
[32] B. Lakshminarayanan, A. Pritzel, and C. Blundell, “Simple and scalable predictive uncertainty estimation using deep ensembles,” in Proc. Adv. Neural Inf. Process. Syst. (NeurIPS), Long Beach, CA, USA, 2017, pp. 6402–6413.
[33] A. Paszke et al., “PyTorch: An imperative style, high-performance deep learning library,” in Proc. Adv. Neural Inf. Process. Syst. (NeurIPS), 2019, pp. 8024–8035.
[34] D. P. Kingma and J. Ba, “Adam: A method for stochastic optimization,” in Proc. 3rd Int. Conf. Learn. Represent. (ICLR), San Diego, CA, USA, 2015.
[35] I. Loshchilov and F. Hutter, “SGDR: Stochastic gradient descent with warm restarts,” in Proc. 5th Int. Conf. Learn. Represent. (ICLR), Toulon, France, 2017.
[36] B. W. Matthews, “Comparison of the predicted and observed secondary structure of T4 phage lysozyme,” Biochim. Biophys. Acta, vol. 405, no. 2, pp. 442–451, 1975.
[37] S. Topal, A. Ercin, C. Kaya, and Y. Gungor, “WasteNet: Waste classification at the edge for smart bins,” arXiv preprint arXiv:2006.05873, 2020.
[38] P. Proença and P. Simões, “TACO: Trash annotations in context for litter detection,” arXiv preprint arXiv:2003.06975, 2020.
[39] G. Hinton, O. Vinyals, and J. Dean, “Distilling the knowledge in a neural network,” arXiv preprint arXiv:1503.02531, 2015.
[40] A. Dosovitskiy et al., “An image is worth 16×16 words: Transformers for image recognition at scale,” in Proc. 9th Int. Conf. Learn. Represent. (ICLR), 2021.
[41] Z. Liu et al., “Swin Transformer: Hierarchical vision transformer using shifted windows,” in Proc. IEEE/CVF Int. Conf. Comput. Vis. (ICCV), Montreal, QC, Canada, 2021, pp. 10012–10022.
How to Cite This Paper
Simran Ghatore, Amritpreet Kaur, Vijay Chhabra (2026). Smart Waste Classification Using Deep Learning for Sustainable Waste Management. International Journal of Computer Techniques, 13(3). ISSN: 2394-2231.
