Electroencephalography (EEG) emotion recognition is critical in human-computer interface applications. EEG is a prevalent tool for discerning various emotional states. However, EEG is confronted with the challenges, including nonstationarity, small amplitude, low signal-to-noise ratio, and significant inter-subject variability. Furthermore, existing methods primarily focus on frequency domain features of EEG signal, overlooking the interchannel relationships. To solve these problems, we introduce a novel approach for emotion recognition using a graph-based conditional distribution-aligned domain adaptation (DA) network. This approach harmonizes both domain-invariant and domain-specific features. With the premise that diverse EEG data share consistent underlying features, we employ dynamic graph convolution to extract the domain-invariant features. Subsequently, we build separate branches for distinct source domains to harness domain-specific features, execute conditional distribution alignment, and eventually determine the scores based on distribution distance. Specifically, the contribution of each source domain is weighted according to its distribution similarity with the target domain, thereby assigning greater importance to more relevant sources. The final outcome is obtained via these weighted scores. The final outcome is ascertained via the associated weighted scores. Based on the leave-one-out cross-validation, we tested our model on the open-source SEED and SEED-IV datasets. The mean accuracy rates were 90.33% and 62.83% in the cross-subject situation, as well as 93.26% and 67.06% in the cross-session situation. Compared with multiple state-of-the-art DA algorithms, our proposed algorithm garners better classification results.

Recently, few-shot strip steel surface defect segmentation has received more and more concerns. However, the existing few-shot segmentation methods usually adopt the frozen encoder, which is pre-trained on the classification task and can only provide class-related knowledge. Therefore, we propose a novel method, namely pre-trained variational auto-encoder based latent gaussian process regression (LGPR), to conduct few-shot strip steel surface defect segmentation. Firstly, different from previous methods, the frozen Variational Auto-Encoder (VAE) based encoder and decoder, which are pre-trained by using the pixel-level self-supervised task (i.e., image reconstruction), can provide rich image-related knowledge. This ensures the effective characterization of defect regions. Secondly, by deploying a gaussian process regression in the latent feature space generated by the VAE-based encoder, pixel-level correlation between support features and query features can be efficiently built. This operation is non-parametric and doesn't bring any training overhead. Besides, we deploy transformer-based projectors to dig long-range contextual cues of support and query features. Extensive experiments are performed on two public datasets, and the experimental results clearly show that our model consistently outperforms the state-of-the-art models with a large margin. Both the codes and results are publicly available at https://github.com/Hlao-hub/LGPR.

Graph Convolutional Networks (GCNs) have shown promise in motor imagery electroencephalogram (EEG) signals classification by modeling spatial dynamics and brain connectivity. However, over-smoothing remains a challenge, leading to homogenized node features and reduced discrimination. To address this, we propose an Adaptive Sparse Awareness-Spatiotemporal Graph Convolutional Network (ASA-STGCN) that combines adaptive sparse graph convolution with attention mechanisms. Notably, a Graph Sparse Convolutional Network (GSCN) in the Adaptive Sparse Awareness Spatial Module (ASAM) enhances brain region feature selection, while the Graph Node Neighborhood Awareness Layer (GNNAL) applies self-attention to reinforce critical topological relationships. The Multi-scale Temporal Convolution Module (MTCM) captures both transient and sustained temporal dependencies. Experimental results achieve accuracies of 97.2%±3.4% (binary) and 83.6%±4.9% (four-class) on BCIC-IV-2a, 96.6%±3.1% (binary) on BCIC-III IVa, and 83.41%±4.3 (binary) on OpenBMI. Discussion confirms the model's effectiveness and its potential to support EEG-based neurorehabilitation and clinical brain computer interface applications.

Motor imagery (MI)-based electroencephalography (EEG) stands as a prominent paradigm in the brain–computer interface (BCI) field, which is frequently applied in neural rehabilitation and gaming due to its accessibility and reliability. Despite extensive research dedicated to MI EEG classification algorithms, a notable deficiency still remains: their performance is often optimal only in subject-specific or dataset-specific scenarios, which undermines their generalization capability, hence restricting BCI systems' practical utility in real-world contexts. To address this limitation, this study introduces a cutting-edge approach: a discriminative adversarial network based on spatial–temporal–graph fusion (STG-DAN). This innovation aims to learn features that are not only class-discriminative but also domain-invariant. Specifically, the feature extraction module guarantees the feature discriminativeness by amalgamating spatial–temporal and graph-related features, while the domain alignment module focuses on both global domain and local subdomain. The two modules are incorporated into one adversarial learning framework to facilitate the acquisition of domain-invariant features. Evaluations on two publicly accessible datasets, BCI competition IV 2a and OpenBMI, affirm the superiority of our proposed model (averaged accuracy = 62.94% and 73.01% for the two datasets in cross-subject circumstance, respectively). In cross-dataset circumstances, it also outperforms several state-of-the-art algorithms, attesting to the potency of STG-DAN.

The performance of modern U-shaped neural networks for medical image segmentation has been significantly enhanced by incorporating Transformer layers. Although Transformer architectures are powerful at extracting global information, its ability to capture local information is limited due to their high complexity. To address this challenge, we proposed a new lightweight U-shaped cascade fusion network (LUCF-Net) for medical image segmentation. It utilized an asymmetrical structural design and incorporated both local and global modules to enhance its capacity for local and global modeling. Additionally, a multi-layer cascade fusion decoding network was designed to further bolster the network's information fusion capabilities. Validation performed on open-source CT, MRI, and dermatology datasets demonstrated that the proposed model outperformed other state-of-the-art methods in handling local-global information, achieving an improvement of 1.46% in Dice coefficient and 2.98 mm in Hausdorff distance on multi-organ segmentation. Furthermore, as a network that combines Convolutional Neural Network and Transformer architectures, it achieves competitive segmentation performance with only 6.93 million parameters and 6.6 gigabytes of floating point operations, without the need for pre-training. In summary, the proposed method demonstrated enhanced performance while retaining a simpler model design compared to other Transformer-based segmentation networks.

The use of transfer learning in brain-computer interfaces (BCIs) has potential applications. As electroencephalogram (EEG) signals vary among different paradigms and subjects, existing EEG transfer learning algorithms mainly focus on the alignment of the original space. They may not discover hidden details owing to the low-dimensional structure of EEG. To effectively transfer data from a source to target domain, a multi-manifold embedding domain adaptive algorithm is proposed for BCI. First, we aligned the EEG covariance matrix in the Riemannian manifold and extracted the characteristics of each source domain in the tangent space to reflect the differences between different source domains. Subsequently, we mapped the extracted characteristics to the Grassmann manifold to obtain a common feature representation. In domain adaptation, the geometric and statistical attributes of EEG data were considered simultaneously, and the target domain divergence matrix was updated with pseudo-labels to maximize the inter-class distance and minimize the intra-class distance. Datasets generated via BCIs were used to verify the effectiveness of the algorithm. Under two experimental paradigms, namely single-source to single-target and multi-source to single-target, the average accuracy of the algorithm on three datasets was 73.31% and 81.02%, respectively, which is more than that of several state-of-the-art EEG cross-domain classification approaches. Our multi-manifold embedded domain adaptive method achieved satisfactory results on EEG transfer learning. The method can achieve effective EEG classification without a same subject's training set.

Based on multiple kernel learning (MKL) support vector machine and decision tree combined strategy, a multi-class classification method is proposed to classify lower limb motions using electromyography (EMG) signals. According to the framework of multiple kernel learning, the MKL-based multi-classifier is constructed using binary tree decomposition method. Four-channel surface EMG signals are firstly collected from lower limb muscles, and then some time-domain features are extracted and inputted into the proposed multi-classifier. Five subdividing patterns are finally identified in level walking, i.e. support prophase, support metaphase, support telophase, swing prophase and swing telophase. The experimental results show that the proposed method can successfully identify these subdividing patterns with better accuracy than standard single-kernel support vector machine classifier.