Traditional human emotion recognition is based on electroencephalogram (EEG) data collection technologies which rely on plenty of rigid electrodes and lack anti-interference, wearing comfort, and portability. Moreover...
详细信息
Traditional human emotion recognition is based on electroencephalogram (EEG) data collection technologies which rely on plenty of rigid electrodes and lack anti-interference, wearing comfort, and portability. Moreover, a significant distribution difference in EEG data also results in low classification accuracy. Here, on-skin biosensors with adhesive and hydrophobic bilayer hydrogel (AHBH) as interfaces for high accuracy emotion classification are proposed. The AHBH achieves remarkable adhesion (59.7 N m(-1)) by combining the adhesion mechanism of catechol groups and electrostatic attraction. Meanwhile, based on the synergistic effects of hydrophobic group rearrangements and surface energy reduction, the AHB-hydrophobic layer exhibits 133.87 degrees water contact angles through hydrophobic treatment of only 0.5 h. Hydrogen and electrostatic bonds are also introduced to form a seamless adhesive-hydrophobic hydrogel interface and inhibit adhesion attenuation, respectively. With the AHBH as an ideal device/skin interface, the biosensor can reliably collect high-quality electrophysiological signals even under vibration, sweating, and long-lasting monitoring condition. Furthermore, the on-skin electrodes, data processing, and wireless modules are integrated into a portable headband for EEG-based emotion classification. A domain adaptive neural network based on the transfer learning technique is introduced to alleviate the effect of domain shift and achieve high classification accuracy.
This article proposes a distributed controller for a network of agents of second-order dynamics to fence a moving target of unknown velocity within their convex hull. Moreover, the agents form a regular polygon format...
详细信息
This article proposes a distributed controller for a network of agents of second-order dynamics to fence a moving target of unknown velocity within their convex hull. Moreover, the agents form a regular polygon formation and avoid collision during the entire evolution. In the control scheme, the agents are not necessarily labeled and the nearest angle rules are applied. Each controller is composed of the functionalities of estimation of the target's velocity, regulation of distance between the agents and the target, and angle repulsion among agents resulting in an equal distribution. The latter two also guarantee collision avoidance. The asymptotical behavior of the closed-loop system is rigorously analyzed, especially subject to the velocity estimation error. The effectiveness of the controller is also demonstrated by numerical simulation.
Rolling bearing is of vital significance in industrial applications and intelligent fault diagnosis (IFD) have been widely exploited in this field. However, cross-machine variations hinder the model performance across...
详细信息
Current few-shot action recognition methods reach impressive performance by learning discriminative features for each video via episodic training and designing various temporal alignment strategies. Nevertheless, they...
详细信息
ISBN:
(数字)9781665469463
ISBN:
(纸本)9781665469463
Current few-shot action recognition methods reach impressive performance by learning discriminative features for each video via episodic training and designing various temporal alignment strategies. Nevertheless, they are limited in that (a) learning individual features without considering the entire task may lose the most relevant information in the current episode, and (b) these alignment strategies may fail in misaligned instances. To overcome the two limitations, we propose a novel Hybrid Relation guided Set Matching (HyRSM) approach that incorporates two key components: hybrid relation module and set matching metric. The purpose of the hybrid relation module is to learn task-specific embeddings by fully exploiting associated relations within and cross videos in an episode. Built upon the task-specific features, we reformulate distance measure between query and support videos as a set matching problem and further design a bidirectional Mean Hausdorff Metric to improve the resilience to misaligned instances. By this means, the proposed HyRSM can be highly informative and flexible to predict query categories under the few-shot settings. We evaluate HyRSM on six challenging benchmarks, and the experimental results show its superiority over the state-of-the-art methods by a convincing margin. Project page: https://***/.
Sparse representation has been widely applied to image classification, where the key issue is to extract a suitable discriminative dictionary. To this end, we propose a joint dictionary and classifier learning algorit...
详细信息
Sparse representation has been widely applied to image classification, where the key issue is to extract a suitable discriminative dictionary. To this end, we propose a joint dictionary and classifier learning algorithm based on a parameterized Bayesian model. Therein, the Gaussian priors of a dictionary endow it with the capability of discrimination and representation. Moreover, we introduce a multivariate Gaussian prior for the sparse codes to achieve group sparsity, thereby substantially improving the classification performance. Furthermore, the sparse codes are estimated by a group-sparse Bayesian learning (GSBL) method, and the dictionary atoms are updated sequentially by maximizing a posterior. Moreover, to avoid manual parameter adjustment, the hyperparameters are optimized by an evidence maximization method. Accordingly, we develop a classification scheme via GSBL. Finally, extensive experiments are conducted on six benchmark datasets of face classification, object recognition, handwritten recognition, and scene categorization to substantiate the effectiveness and superiority of the proposed method.
Most existing wideband signal detection and recognition (WSDR) methods rely on diverse, large-scale, and well-labeled training data, which are often difficult to obtain in practical application scenarios such as non-c...
详细信息
Most existing wideband signal detection and recognition (WSDR) methods rely on diverse, large-scale, and well-labeled training data, which are often difficult to obtain in practical application scenarios such as non-cooperative environments and novel signaling regimes. In this article, we propose a method for constructing a virtual signal large model (VSLM) and applying it to tackle the WSDR challenge under few-shot or even cross-domain few-shot scenarios. Firstly, we design two plug-and-play modules, virtual sample generation (VSG) and virtual category generation (VCG), for VSLM, respectively. VSG simulates the local and overall relationship between the burst signal and the constant signal, which is mainly completed by extracting time-frequency meta-block and data enhancement. Based on VSG and the multi-label concept, we further create virtual novel categories by injecting customizable semantic information into meta-blocks. Then, we further propose a dual decoupled network (DDN) to train the VSLM. DDN enhances signal details by decoupling low gray values (DLGV) in time-frequency representation, and alleviates conflicts during multi-task joint optimization by decoupling spectrum localization and signal classification. Finally, based on the wideband spectrogram dataset, extensive experiments have validated that our proposed methods can significantly improve the performance of WSDR under few-shot conditions.
Since the fixed-time stability forms of nonlinear systems satisfy strict conditions, there are few general forms for nonlinear systems to achieve fixed-time stability. This work proposes a new class of more general fi...
详细信息
Since the fixed-time stability forms of nonlinear systems satisfy strict conditions, there are few general forms for nonlinear systems to achieve fixed-time stability. This work proposes a new class of more general fixed-time stability criteria. It is worth mentioning that, compared with the traditional method of estimating the convergence time, this paper obtains a more conservative stable time estimation formula through the integration method of the generalized integral mean theorem. In addition, given that the fixed-time stabilization of neural networks and chaotic oscillators have attracted extensive attention in recent years, and there are still many fixed-time stabilizations of nonlinear systems that have not been studied. Therefore, a discontinuous controller is designed in this paper. The above stability theory results are applied to the fixed-time stabilization of the Takagi-Sugeno (T-S) fuzzy competitive neural network and chaotic system (coupled Chua's oscillator). Finally, the validity and applicability of the theoretical results are verified by examples.
In the field of industrial, robots are becoming a modern method to perform automatic washing on large industrial components while detecting dirt. Therefore, the detection and segmentation of dirt have a great impact o...
详细信息
ISBN:
(纸本)9783031189128;9783031189135
In the field of industrial, robots are becoming a modern method to perform automatic washing on large industrial components while detecting dirt. Therefore, the detection and segmentation of dirt have a great impact on optimizing the effects and improving quality of washing by modern washing robots. We propose DDSN (Dirt Detection and Segmentation Network) in this paper which improves the SVDD (Support Vector Data Description), using neural networks to obtain the optimal kernel function of SVDD, so that training images with no dirt can be mapped to the smallest hypersphere in the feature space. In dirt detection and segmentation, the distance between the test images and the centers of the corresponding hypersphere is defined as the feature value, and the dirt scores of the pixels can be obtained. Before training of dirt dataset, we use a larger anomaly detection dataset named MVTecAD which is also the one-class classification to pretrain the feature extraction network, which makes up for the lack of samples in the dirt dataset and speeds up the convergence of the model. Afterwards, we transfer the feature extraction network to training the dirt dataset of large industrial components. The results show that the methods proposed in this paper performs well in detection and segmentation of both MVTecAD dataset and dirt dataset of large industrial components.
Partial occlusion effects are a phenomenon that blurry objects near a camera are semi-transparent, resulting in partial appearance of occluded background. However, it is challenging for existing bokeh rendering method...
详细信息
ISBN:
(纸本)9783031200670;9783031200687
Partial occlusion effects are a phenomenon that blurry objects near a camera are semi-transparent, resulting in partial appearance of occluded background. However, it is challenging for existing bokeh rendering methods to simulate realistic partial occlusion effects due to the missing information of the occluded area in an all-in-focus image. Inspired by the learnable 3D scene representation, Multiplane image (MPI), we attempt to address the partial occlusion by introducing a novel MPI-based high-resolution bokeh rendering framework, termed MPIB. To this end, we first present an analysis on how to apply the MPI representation to bokeh rendering. Based on this analysis, we propose an MPI representation module combined with a background inpainting module to implement high-resolution scene representation. This representation can then be reused to render various bokeh effects according to the controlling parameters. To train and test our model, we also design a ray-tracing-based bokeh generator for data generation. Extensive experiments on synthesized and real-world images validate the effectiveness and flexibility of this framework.
In this paper, a sensor network is used to implement the distributed observer design problem for a linear time-invariant mobile target system with constant or time-varying velocity. Each sensor can only access a part ...
详细信息
In this paper, a sensor network is used to implement the distributed observer design problem for a linear time-invariant mobile target system with constant or time-varying velocity. Each sensor can only access a part of the output information of the target. Two types of interacting dynamics are endowed to each sensor, one is a consensus-based algorithm for the state estimation using the local measurements, while the other is a leader-following flocking-like algorithm such that the mobile sensors can avoid collision, maintain communication, and track the target. By adopting adaptive coupling gains on the consensus-based term in the state estimation algorithm, a fully distributed observer which is independent of the communication topology associated with the sensor network is constructed. Finally, numerical simulations are presented to illustrate the effectiveness of the proposed theoretical results.
暂无评论