Self‐training maximum classifier discrepancy for EEG emotion recognition

Even with an unprecedented breakthrough of deep learning in electroencephalography(EEG),collecting adequate labelled samples is a critical problem due to laborious and time‐consuming labelling.Recent study proposed to solve the limited label problem via domain adaptation methods.However,they mainly focus on reducing domain discrepancy without considering task‐specific decision boundaries,which may lead to feature distribution overmatching and therefore make it hard to match within a large domain gap completely.A novel self‐training maximum classifier discrepancy method for EEG classification is proposed in this study.The proposed approach detects samples from a new subject beyond the support of the existing source subjects by maximising the discrepancies between two classifiers'outputs.Besides,a self‐training method that uses unlabelled test data to fully use knowledge from the new subject and further reduce the domain gap is proposed.Finally,a 3D Cube that incorporates the spatial and frequency information of the EEG data to create input features of a Convolutional Neural Network(CNN)is constructed.Extensive experiments on SEED and SEED‐IV are conducted.The experimental evaluations exhibit that the proposed method can effectively deal with domain transfer problems and achieve better performance.

Optimizing depth of field in 3D light-field display by analyzing and controlling light-beam divergence angle

A concept of divergence angle of light beams(DALB)is proposed to analyze the depth of field(DOF)of a 3D light-field display system.The mathematical model between DOF and DALB is established,and the conclusion that DOF and DALB are inversely proportional is drawn.To reduce DALB and generate clear depth perception,a triple composite aspheric lens structure with a viewing angle of 100°is designed and experimentally demonstrated.The DALB-constrained 3D light-field display system significantly improves the clarity of 3D images and also performs well in imaging at a 3D scene with a DOF over 30 cm.

Intelligent Detection and Identification in Fiber-Optical Perimeter Intrusion Monitoring System Based on the FBG Sensor Network

A real-time intelligent fiber-optic perimeter intrusion detection system （PIDS） based on the fiber Bragg grating （FBG） sensor network is presented in this paper. To distinguish the effects of different intrusion events, a novel real-time behavior impact classification method is proposed based on the essential statistical characteristics of signal＇s profile in the time domain. The features are extracted by the principal component analysis （PCA）, which are then used to identify the event with a K-nearest neighbor classifier. Simulation and field tests are both carried out to validate its effectiveness. The average identification rate （IR） for five sample signals in the simulation test is as high as 96.67%, and the recognition rate for eight typical signals in the field test can also be achieved up to 96.52%, which includes both the fence-mounted and the ground-buried sensing signals. Besides, critically high detection rate （DR） and low false alarm rate （FAR） can be simultaneously obtained based on the autocorrelation characteristics analysis and a hierarchical detection and identification flow.