A transfer learning-based convolutional neural network and its novel application in ship spare-parts classification

Yang, Ke; Yang, Taiwei; Yao, Yunan; Fan, Shidong

doi:10.1016/j.ocecoaman.2021.105971

Cited by 11 publications

(4 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This is different from the results obtained by applying the ImagetNet, which indicated that the deeper the VGGNet structures, the better the performance. This phenomenon can be ascribed to the dataset of plastic that is not as complex as those from the ImageNet (Yang et al., 2021). These results suggested that the choice of structure and depth should be appropriate to the practical application.…”

Section: Discussionmentioning

confidence: 99%

Data‐driven models employed to waste plastic in China: Generation, classification, and environmental assessment

Lin

Zhao

Zhang

et al. 2022

J of Industrial Ecology

View full text Add to dashboard Cite

It is crucial to precisely predict the generation of plastic waste and realize its fine classification in terms of policy demand and environmental benefits. Only a few studies have used the date‐driven model to forecast the amount of plastic waste. The benefits of classifying the plastic waste using deep learning have also been only scarcely reported in the literature. Therefore, this study used the Prophet model to estimate the amount of plastic waste from 2005 to 2025. Four types of Visual Geometry Group Networks based on Transfer Learning (TLVGGNet) were performed for classifying the plastic waste. Potentials of saved energy, the reduction of green‐house gases emission (GHG), and air pollutants were also discussed under different scenarios (current recycling system and TLVGGNet system). The results showed that the amount of waste plastic was anticipated to be 26.44 Mt in 2020 and 33.18 Mt in 2025 in China. The method of transfer learning could shorten the training time and improve the performance of the TLVGGNet‐11 model in the test dataset (41.6–68.1%). Moreover, TLVGGNet‐16 was considered to be the most optimum model for plastic waste classification in terms of training time (83.94 s), accuracy (75.5%), precision (76.9%), recall (75.5%), and F1 score (75.1%). The TLVGGNet‐16 system contributes about 12.15–15.97% in terms of electricity‐savings. Compared with the current recycling system, the amount of CO2 emissions saved and reduction in CH4 emissions could be more than 8–10% and 0.4–0.5%, respectively, in the TLVGGNet‐16 system. The saved VOCS and NOX emissions were within the ranges of 34.84–127.22 billion kg and 93.64–414.14 billion kg between 2017 and 2025 using the method of deep learning.

show abstract

Section: Discussionmentioning

confidence: 99%

Data‐driven models employed to waste plastic in China: Generation, classification, and environmental assessment

Lin

Zhao

Zhang

et al. 2022

J of Industrial Ecology

View full text Add to dashboard Cite

show abstract

“…This is different from the conclusion of the ImageNet application, which indicated that the deeper ResNet structures, the better performance. This phenomenon can be ascribed to the dataset of C&D waste being not as complex as these from the ImageNet (Yang et al, 2021). Those results suggested that the structure and depth choice of C&DWNet models should be made according to the practical application.…”

Section: Rocmentioning

confidence: 94%

Applying machine learning to fine classify construction and demolition waste based on deep residual network and knowledge transfer

Lin

Zhao

Ting-ting

et al. 2022

Environ Dev Sustain

View full text Add to dashboard Cite

Few studies reported using the convolutional neural network with transfer learning to finely classify the construction and demolition waste. Objectives: This study aims to develop a highly efficient method to realize the finely sorting the construction and demolition waste, which is a key step for promoting the recycling system to realize carbon neutrality in the waste management sector. Methodology: C&DWNet models, ResNet structures based on knowledge transfer and cyclical learning rate, were proposed to classify ten types of construction and demolition waste. Indexes (confusion metric, accuracy, precision, recall, F1 score, sensitivity, specificity and kappa) were adopted to evaluate the performance of various C&DWNet models. Results: Knowledge transfer can reduce the training time and improve the performance of the C&DWNet model. The average training time is increased with the increase of the layer of C&DWNet architecture from C&DWNet-18 (946.7 s) to C&DWNet-152 (1186.6 s). The accuracy of various C&DWNet models is approximately 72~74%, the best accuracy is 73.6% in C&DWNet-152. C&DWNet-18 is more suitable for the classification of construction and demolition waste in terms of training time, accuracy, precision, and F1 score. Moreover, the t-distributed stochastic neighbor embedding can distinctly separate each type of construction and demolition waste. Improvement: The environmental applications and limitations of the C&DWNet module were also discussed, which could provide a reference for the intelligent management of construction and demolition waste and promote the development of the circular economy.

show abstract

“…Yang et al classified spare parts into three categories using TL [8]. They used the VGG19, Alexnet, and ResNet50 models, with the VGG19 model yielding the highest success rate at 0.963.…”

Section: Related Workmentioning

confidence: 99%

“…In [4], implementing the TL-FT2 scenario, the accuracy of VGG16 is 0.9146. In [8], implementing the TL-FT2 scenario, the accuracy of ResNet50 is 0.8983. In [10], implementing the TL-FT2 scenario, the accuracy of Xception is 0.874.…”

Section: Creating Datasetmentioning

confidence: 99%

A Hybrid Classification Approach for Fasteners Based on Transfer Learning with Fine-Tuning and Deep Features

TAŞTİMUR,

AKIN

2023

Turkish Journal of Science and Technology

View full text Add to dashboard Cite

Deep learning, which has seen frequent use in recent studies, has helped solve the problem of classifying objects of many different types and properties. Most studies both create and train a convolutional neural network (CNN) from scratch. The time spent training the network is thus wasted. Transfer learning (TL) is used both to prevent the loss of time due to training the dataset and to more effectively classify small datasets. This study performs classification using a dataset containing eighteen types of fastener. Our study contains three different TL scenarios. Two of them use TL with fine-tuning (FT), while the third does so with feature extraction (FE). The study compares the classification performance of eighteen different pre-trained network models (i.e., one or more versions of EfficientNet, DenseNet, InceptionResNetV2, InceptionV3, MobileNet, ResNet50, Xception, and VGGNet) in detail. When compared to other research in the literature, our first and second scenarios provide excellent implementations of TL-FT, while our third scenario, TL-FE, is hybrid and produces better results than the other two. Furthermore, our findings are superior to those of most previous studies. The models with the best results are DenseNet169 with an accuracy of 0.97 in the TL-FT1 scenario, EfficientNetB0 with 0.96 in TL-FT2, and DenseNet169 with 0.995 in TL-FE.

show abstract

A transfer learning-based convolutional neural network and its novel application in ship spare-parts classification

Cited by 11 publications

References 21 publications

Data‐driven models employed to waste plastic in China: Generation, classification, and environmental assessment

Data‐driven models employed to waste plastic in China: Generation, classification, and environmental assessment

Applying machine learning to fine classify construction and demolition waste based on deep residual network and knowledge transfer

A Hybrid Classification Approach for Fasteners Based on Transfer Learning with Fine-Tuning and Deep Features

Contact Info

Product

Resources

About