Journal article
Enhanced transfer learning with ImageNet trained classification layer
Image and Video Technology, Vol.11854, pp.142-155
2019
Abstract
Parameter fine tuning is a transfer learning approach whereby learned parameters from pre-trained source network are transferred to the target network followed by fine-tuning. Prior research has shown that this approach is capable of improving task performance. However, the impact of the ImageNet pre-trained classification layer in parameter fine-tuning is mostly unexplored in the literature. In this paper, we propose a fine-tuning approach with the pre-trained classification layer. We employ layer-wise fine-tuning to determine which layers should be frozen for optimal performance. Our empirical analysis demonstrates that the proposed fine-tuning performs better than traditional fine-tuning. This finding indicates that the pre-trained classification layer holds less category-specific or more global information than believed earlier. Thus, we hypothesize that the presence of this layer is crucial for growing network depth to adapt better to a new task. Our study manifests that careful normalization and scaling are essential for creating harmony between the pre-trained and new layers for target domain adaptation. We evaluate the proposed depth augmented networks for fine-tuning on several challenging benchmark datasets and show that they can achieve higher classification accuracy than contemporary transfer learning approaches.
Details
- Title
- Enhanced transfer learning with ImageNet trained classification layer
- Authors/Creators
- T. Shermin (Author/Creator) - Federation UniversityS.W. Teng (Author/Creator) - Federation UniversityM. Murshed (Author/Creator) - Federation UniversityG. Lu (Author/Creator) - Federation UniversityF. Sohel (Author/Creator) - Murdoch UniversityM. Paul (Author/Creator) - Charles Sturt University
- Publication Details
- Image and Video Technology, Vol.11854, pp.142-155
- Publisher
- Springer Verlag
- Identifiers
- 991005541083307891
- Copyright
- © 2019 Springer Nature Switzerland AG
- Murdoch Affiliation
- Information Technology, Mathematics and Statistics
- Language
- English
- Resource Type
- Journal article
- Additional Information
- Conference paper from Pacific-Rim Symposium on Image and Video Technology (PSIVT) 2019; Sydney, NSW. 18 - 22 November 2019
UN Sustainable Development Goals (SDGs)
This output has contributed to the advancement of the following goals:
Source: InCites
Metrics
28 Record Views