返回
A multi-label waste detection model based on transfer learning
DOI:10.1016/j.resconrec.2022.106235.png)
摘要
En 中文
Accurate and efficient treatment of domestic waste is an important part of urban management. Whether domestic waste can be classified effectively will affect the sustainable development of human society. Previous research on the problem of waste image classification has focused on single-category waste recognition, which falls short of meeting the needs of real waste classification scenarios. In this study, a YOLO-WASTE multi-label waste clas-sification model based on transfer learning is constructed to realize the fast recognition and classification of multiple wastes. To speed up and optimize the learning efficiency of the model, a multi-label waste image dataset is also created, with each image including multiple wastes or multiple categories of waste. The experimental result shows that the mAP value of the YOLO-WASTE model is 92.23%, and the average time of detecting an image is 0.424 s, its classification performance is significantly better than other image classification algorithms. The proposed YOLO-WASTE model provides new insights into complex waste identification and has the potential to help advance efficient waste management for sustainable urban development.
Keyword:
Transfer learning
Image recognition
Recyclable waste classification
Deep learning
Multi-target detection
YOLOv4
期刊
R
IF:
10.9
论文数:
7.1K
被引数:
5.3W
机构
引用论文
Using channel pruning-based YOLO v4 deep learning algorithm for the real-time and accurate detection of apple flowers in natural environments基于通道剪枝的YOLO v4深度学习算法在自然环境下苹果花朵实时准确检测中的应用
Synthesis of n-type semiconducting diamond film using diphosphorus pentaoxide as the doping source以五氧化二磷为掺杂源合成n型半导体金刚石膜
Deep learning in material recovery: Development of method to create training database材料恢复中的深度学习: 创建训练数据库的方法开发
Structural study of lanthanides(III) in aqueous nitrate and chloride solutions by EXAFS通过EXAFS对硝酸盐和氯化物水溶液中镧系元素 (III) 的结构研究

