Research on Intelligent Waste Classification Detection Based on an Enhanced ResNet50

View at publisher Summarise with VIRAT
Authors: Chenlu Guo
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

This study proposes an intelligent waste classification system based on an enhanced ResNet50 architecture, aimed at addressing the challenges of automated waste sorting in urban environments. We present a deep learning framework that attains superior classification performance by integrating a Squeeze-and-Excitation (SE) attention mechanism into the ResNet50 backbone and employing an optimized training strategy. The study utilizes a diverse waste dataset comprising four primary categories: recyclables, kitchen (food) waste, hazardous waste, and other (residual) waste. To address class imbalance and improve model robustness, we implemented the Focal Loss function and the AdamW optimizer with a cosine learning-rate schedule. Extensive comparative experiments demonstrate the superiority of our enhanced model. The system achieved a 99.5% test accuracy and an F1 score of 0.9950, significantly outperforming baseline models such as the original ResNet50 and VGG16. The results confirm that our approach offers notable advantages in both accuracy and computational efficiency, making it well-suited for practical deployment within smart city infrastructure and environmental management systems.

Fields of Study & Subjects

Advanced Neural Network Applications Municipal Solid Waste Management Waste Management and Recycling

Publication details

Journal / source
IEEE International Conference on Power, Electronics and Green Energy (ICPEGE)
Pages
1771–1775
Publication date
28 January 2026
Language
EN