Convolutional Neural Network (CNN)-based image recognition is an approach to achieving intelligent classification of aerospace foods. We propose an image classification methodology based on deep migration learning to increase the accuracy and timeliness of aerospace food image identification. Firstly, the aerospace food images are acquired, and the Aerospacefood-9 dataset is obtained by data enhancement. Then, a portion of CNN model is frozen while the model training is completed by fine-tuning weights utilizing migration learning. Four models, VGG16, ResNet50, MobileNetV3, and EfficientNetV2, are selected for these experiments. The results show that the deep migration models improve the classification accuracy and timeliness of aerospace food images. Among them, the MobileNetV3 model works best and achieves 99.6% classification accuracy on the validation set.
Deep Migration Learning-Based Aerospace Food Image Classification
Lect. Notes Electrical Eng.
International Conference on Man-Machine-Environment System Engineering ; 2023 ; Beijing, China October 20, 2023 - October 23, 2023
2023-09-05
8 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch
Secure Image Classification Using Deep Learning
Springer Verlag | 2023
|DEEP LEARNING DEVELOPMENT PLATFORM FOR AEROSPACE APPLICATIONS
TIBKAT | 2021
|NTRS | 1970
NTRS | 1990
|