Study on the CNN model optimization for household garbage classification based on machine learning
Wenzhuo Xie,Shiping Li,Wei Xu,Haotian Deng,Weihan Liao,Xianbao Duan,Xuehua Wang
DOI: https://doi.org/10.3233/ais-220017
2022-11-19
Journal of Ambient Intelligence and Smart Environments
Abstract:In order to solve the problem of household garbage classification accurately and efficiently, convolutional neural network classifier is an effective method. In this study, a garbage classification device was designed, and the image dataset Wit-Garbage for garbage classification was constructed based on the device by collecting garbage images under different light intensity and weather environment. The performances of the five network models VGG16, ResNet50, DenseNet121, MobileNet V2, Inception V3 on this dataset were compared by transfer learning. Then, the lightweight convolutional neural network MobileNet V2 was optimized by fine-tuning the hyperparameters, such as the type of optimizer, learning rate, Dropout parameter and number of freezing layers, respectively, and the training accuracy and efficiency were discussed in detail. Finally, the optimized model MobileNet V2 was deployed to the self-made garbage classification device for verification. The results show that the MobileNet V2 network model is superior to other networks in terms of training accuracy and efficiency on the proposed dataset, when the image input size was 224 ∗ 224 pixels, the Adamax optimizer was adopted, the learning rate was 0.0001, the Dropout was less than 0.5, and the number of frozen layers is less than 30. The actual verification results show that the average accuracy of the optimized network model trained on the proposed dataset for MSW classification was up to 98.75%, and compared with the model before optimization, the average accuracy was improved by 2.83%, and the average detection time was reduced by 69%.
computer science, information systems,telecommunications, artificial intelligence