Related Experiment Video
Updated: Jul 8, 2025
03:31
Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
Published on: December 15, 2023
558
GCDN-Net: Garbage classifier deep neural network for recyclable urban waste management
Md Mosarrof Hossen1, Azad Ashraf2, Mazhar Hasan2
1Department of Electrical and Electronics Engineering, University of Dhaka, Dhaka, Bangladesh.
Waste Management (New York, N.Y.)
|December 19, 2023
Summary
Advanced deep learning models, like the Garbage Classifier Deep Neural Network (GCDN-Net), can accurately sort waste images. This technology offers a promising solution for automated waste management and recycling systems.
Area of Science:
- Environmental Science
- Computer Science
- Artificial Intelligence
Background:
- Increasing urbanization and population growth necessitate advanced waste management solutions.
- Deep learning models show potential for automated waste sorting due to their image recognition capabilities.
Purpose of the Study:
- To propose a multi-stage machine learning approach for classifying diverse waste categories.
- To introduce and evaluate the novel Garbage Classifier Deep Neural Network (GCDN-Net) for waste image classification.
Main Methods:
- Utilized the "Garbage In, Garbage Out" (GIGO) dataset comprising 25,000 waste images.
- Developed and implemented the Garbage Classifier Deep Neural Network (GCDN-Net) for both single-label and multi-label classification.
- Evaluated GCDN-Net performance against state-of-the-art waste classification methods and used Score-CAM for visualization.
Main Results:
- GCDN-Net achieved 95.77% accuracy, 95.78% precision, 95.77% recall, and 95.77% F1-score in single-label classification, outperforming existing models.
- In multi-label classification, GCDN-Net attained a Mean Average Precision (mAP) of 0.69 and an F1-score of 75.01%.
Conclusions:
- Deep learning models, particularly GCDN-Net, are effective for categorizing diverse waste types.
- Automated waste sorting systems powered by deep learning can reduce costs and processing times in waste management.
