Journal
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
Volume -, Issue -, Pages -Publisher
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TNNLS.2022.3185795
Keywords
Training; Task analysis; Testing; Feature extraction; Hyperspectral imaging; Power capacitors; Electronic mail; Cross-scene; distribution alignment; domain adaption; few-shot learning (FSL); graph neural network (GNN); hyperspectral image classification
Categories
Funding
- National Natural Science Foundation of China [61922013, U1833203, 62001023]
- Beijing Natural Science Foundation [JQ20021, L191004]
- China Postdoctoral Science Foundation [BX20200058]
Ask authors/readers for more resources
This study proposes a new domain adaptation method that combines few-shot learning with domain alignment based on graph information aggregation to address the issue of reduced classification performance in the presence of new classes in target data. By training on source data with all label samples and target data with a few label samples, two key blocks are designed to extract and compare features and distributions across different domains, and cross-domain graph alignment methods are used to mitigate the impact of domain shift on few-shot learning. Experimental results demonstrate the superiority of the proposed method.
Most domain adaptation (DA) methods in cross-scene hyperspectral image classification focus on cases where source data (SD) and target data (TD) with the same classes are obtained by the same sensor. However, the classification performance is significantly reduced when there are new classes in TD. In addition, domain alignment, as one of the main approaches in DA, is carried out based on local spatial information, rarely taking into account nonlocal spatial information (nonlocal relationships) with strong correspondence. A graph information aggregation cross-domain few-shot learning (Gia-CFSL) framework is proposed, intending to make up for the above-mentioned shortcomings by combining FSL with domain alignment based on graph information aggregation. SD with all label samples and TD with a few label samples are implemented for FSL episodic training. Meanwhile, intradomain distribution extraction block (IDE-block) and cross-domain similarity aware block (CSA-block) are designed. The IDE-block is used to characterize and aggregate the intradomain nonlocal relationships and the interdomain feature and distribution similarities are captured in the CSA-block. Furthermore, feature-level and distribution-level cross-domain graph alignments are used to mitigate the impact of domain shift on FSL. Experimental results on three public HSI datasets demonstrate the superiority of the proposed method. The codes will be available from the website: https://github.com/YuxiangZhang-BIT/IEEE_TNNLS_Gia-CFSL.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available