Abstract:In the application of brain-computer interface, the differences in imaging methods and brain structure between subjects hinder the effectiveness of decoding algorithms when applied on different subjects. Transfer learning has been designed to solve this problem. There have been many applications of transfer learning in motor imagery (MI), however the effectiveness is still limited due to the inconsistent domain alignment, lack of prominent data features and allocation of weights in trails. In this paper, a Multi-layer transfer learning algorithm based on improved Common Spatial Patterns (MTICSP) was proposed to solve these problems. Firstly, the source domain data and target domain data were aligned by Target Alignment (TA)method to reduce distribution differences between subjects. Secondly, the mean covariance matrix of the two classes was re-weighted by calculating the distance between the covariance matrix of each trial in the source domain and the target domain. Thirdly, the improved Common Spatial Patterns (CSP) by introducing regularization coefficient was proposed to further reduce the difference between source domain and target domain to extract features. Finally, the feature blocks of the source domain and target domain were aligned again by Joint Distribution Adaptation (JDA) method. Experiments on two public datasets in two transfer paradigms multi-source to single-target (MTS) and single-source to single-target (STS) verified the effectiveness of our proposed method. The MTS and STS in the 5-person dataset were 80.21% and 77.58%, respectively, and 80.10% and 73.91%, respectively, in the 9-person dataset. Experimental results also showed that the proposed algorithm was superior to other state-of-the-art algorithms. In addition, the generalization ability of our algorithm MTICSP was validated on the fatigue EEG dataset collected by ourselves, and obtained 94.83% and 87.41% accuracy in MTS and STS experiments respectively. The proposed method combines improved CSP with transfer learning to extract the features of source and target domains effectively, providing a new method for combining transfer learning with motor imagination.

Deep Sparse Informative Transfer SoftMax for Cross-Domain Image Classification.

Cross-domain Residual Deep NMF for Transfer Learning Between Different Hyperspectral Image Scenes.

Cross-Domain Hyperspectral Image Classification Based on Graph Convolutional Networks

Cross-Domain Hyperspectral Image Classification Based on Transformer

Domain‐invariant Attention Network for Transfer Learning Between Cross‐scene Hyperspectral Images

Building Cross-Domain Mapping Chains from Multi-CycleGAN for Hyperspectral Image Classification

Cross-domain attention network for hyperspectral image classification

Sparse Deep Transfer Learning for Convolutional Neural Network

Learning a Deep Structural Subspace Across Hyperspectral Scenes with Cross-Domain VAE

Cross-city Landuse classification of remote sensing images via deep transfer learning

Cross-Modality Bridging and Knowledge Transferring for Image Understanding

Improving the Generalization Ability of Deep Neural Networks for Cross-Domain Visual Recognition

An Informative Logistic Regression for Cross-Domain Image Classification

Transfer Classification for Distinct Manifestations with Shared Information

Discriminative Sparse Filtering for Multi-Source Image Classification

Experiments in cross-domain few-shot learning for image classification

Deep Transfer Network: Unsupervised Domain Adaptation

Deep Subdomain Adaptation Network for Image Classification

Multi-layer transfer learning algorithm based on improved common spatial pattern for brain-computer interfaces

Nonconvex and discriminative transfer subspace learning for unsupervised domain adaptation

Multi-domain few-shot image recognition with knowledge transfer