A Brain-Computer Interface (BCI) is designed for human-computer interactions without body movement. To improve the expression of input features closely related to a given BCI task, we propose a convolution network called GANet to analyze Event-Related Potential (ERP) in the BCI task. This model introduces the parallel convolution to extract multi-scale features in electroencephalogram (EEG) data. In addition, a Graph-based Attention (GA) mechanism is used to model interdependencies among different EEG channels. Experiments are conducted on a public dataset of 15 subjects in the specific-subject and cross-subject scenarios. The results indicate that the GANet achieves state-of-the-art performance with an accuracy of 99.75% in the specific-subject scenario and an accuracy of 81.37% in the cross-subject scenario. Different structures are discussed to analyze the contributions of the parallel convolution in feature extraction and the GA module in feature expression. GANet shows satisfied performance and good generalization in the BCI task. Our codes are publicly available at https://github.com/Debbie-85/GANet.

Detecting and treating older adults who fall in an environment without others is essential. Millimeter-wave radar sensors do not have the disadvantage of invading user privacy like cameras, nor do they require users to wear them in real-time like wearable devices. Actual samples of older adults fall are difficult to collect, and it is unethical to require older adults to fall repeatedly to collect data. In addition, different body types and action patterns will inevitably reduce the model’s performance when new users use the model. In this paper, we constructed a fall detection model based on anomaly detection. The model is trained only using non-fall samples and detects falls as abnormal actions. The proposed model uses a domain generalization architecture based on domain feature alignment to extract domain-invariant features of the model, thereby improving the model’s generalization ability. In addition, we introduced the idea of denoising learning into the feature extractor and feature predictor to improve the model’s anti-interference ability. We conducted sufficient experiments to explore the effectiveness of the proposed method. When tested with new domain data, the proposed model has a true positive rate of 96.12%, a false positive rate of 0.97%, and an area under the receiver operating characteristic of 0.9979.