Version 1
: Received: 2 May 2023 / Approved: 3 May 2023 / Online: 3 May 2023 (05:56:00 CEST)
Version 2
: Received: 30 May 2023 / Approved: 1 June 2023 / Online: 1 June 2023 (04:39:38 CEST)
Zhou, B.; Zhao, J.; Yan, C.; Zhang, X.; Gu, J. Global and Local Knowledge Distillation Method for Few-Shot Classification of Electrical Equipment. Appl. Sci.2023, 13, 7016.
Zhou, B.; Zhao, J.; Yan, C.; Zhang, X.; Gu, J. Global and Local Knowledge Distillation Method for Few-Shot Classification of Electrical Equipment. Appl. Sci. 2023, 13, 7016.
Zhou, B.; Zhao, J.; Yan, C.; Zhang, X.; Gu, J. Global and Local Knowledge Distillation Method for Few-Shot Classification of Electrical Equipment. Appl. Sci.2023, 13, 7016.
Zhou, B.; Zhao, J.; Yan, C.; Zhang, X.; Gu, J. Global and Local Knowledge Distillation Method for Few-Shot Classification of Electrical Equipment. Appl. Sci. 2023, 13, 7016.
Abstract
With the increasing use of intelligent mobile devices for online inspection of electrical equipment in smart grids, the limited computing power and storage of these devices pose challenges for carrying large algorithm models and it’s hard to obtain a large number of images of electrical equipment in public. In this paper, we propose a novel distillation method that compresses the knowledge of teacher networks into a small few-shot classification network using a global and local knowledge distillation strategy. Central to our method is exploiting the global and local relationship between the features exacted by the backbone of the teacher network and student network. We compare our method with recent state-of-the-art methods in three public datasets and achieve the best performance. We also contribute a new dataset, EEI-100, specifically designed for classification of electrical equipment, and demonstrate that our method achieves a prediction accuracy of 94.12% with only 5-shot images.
Computer Science and Mathematics, Artificial Intelligence and Machine Learning
Copyright:
This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.