Kai Wang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Kai Wang has a total of 9 co-authors. Here are the top co-authors by number of shared publications:
Ding-Hua Zhang
Mechanical and Electronic Engineering Institute, Northwestern Polytechnical University, China
1 shared publication- Hanying Yan
Department of Biostatistics, Epidemiology and Informatics, University of Pennsylvania Perelman School of Medicine, Philadelphia (H.Y., M.L.).
1 shared publication - Jianhua Zou
School of Automation Science and Engineering, Faculty of Electronic and Information Engineering, Xian Jiaotong University, China
1 shared publication - Kuidong Huang
Northwestern Polytechnical Univ. (China)
1 shared publication - Ling Zou
School of Computer Science and Artificial Intelligence, Changzhou University, Changzhou, China
1 shared publication - Wanshun Gao
School of Electronic and Information Engineering, Xi’an Jiaotong University, Xi’an 710049, China
1 shared publication
Latest Publications
- 25 February 2019
- Vol. 36 (1) , 7-15
[Study of denoising of simultaneous electroencephalogram-functional magnetic resonance imaging signal based on real-time constrained independent components analysis].
Simultaneous recording of electroencephalogram (EEG)-functional magnetic resonance imaging (fMRI) plays an important role in scientific research and clinical field due to its high spatial and temporal resolution. However, the fusion results are seriously influenced by ballistocardiogram (BCG) artifacts under MRI environment. In this paper, we improve the off-line constrained independent components analysis using real-time technique (rt-cICA), which is applied to the simulated and real resting-state EEG data. The results show that for simulated data analysis, the value of error in signal amplitude (Er) obtained by rt-cICA method was obviously lower than the traditional methods such as average artifact subtraction ( P<0.005). In real EEG data analysis, the improvement of normalized power spectrum (INPS) calculated by rt-cICA method was much higher than other methods ( P<0.005). In conclusion, the novel method proposed by this paper lays the technical foundation for further research on the fusion model of EEG-fMRI.
- 25 February 2019
- Vol. 36 (1) , 7-15
[Study of denoising of simultaneous electroencephalogram-functional magnetic resonance imaging signal based on real-time constrained independent components analysis].
Simultaneous recording of electroencephalogram (EEG)-functional magnetic resonance imaging (fMRI) plays an important role in scientific research and clinical field due to its high spatial and temporal resolution. However, the fusion results are seriously influenced by ballistocardiogram (BCG) artifacts under MRI environment. In this paper, we improve the off-line constrained independent components analysis using real-time technique (rt-cICA), which is applied to the simulated and real resting-state EEG data. The results show that for simulated data analysis, the value of error in signal amplitude (Er) obtained by rt-cICA method was obviously lower than the traditional methods such as average artifact subtraction ( P<0.005). In real EEG data analysis, the improvement of normalized power spectrum (INPS) calculated by rt-cICA method was much higher than other methods ( P<0.005). In conclusion, the novel method proposed by this paper lays the technical foundation for further research on the fusion model of EEG-fMRI.
- Open Access
A Coarse-to-Fine Approach for 3D Facial Landmarking by Using Deep Feature Fusion
Facial landmarking locates the key facial feature points on facial data, which provides not only information on semantic facial structures, but also prior knowledge for other kinds of facial analysis. However, most of the existing works still focus on the 2D facial image which may suffer from lighting condition variations. In order to address this limitation, this paper presents a coarse-to-fine approach to accurately and automatically locate the facial landmarks by using deep feature fusion on 3D facial geometry data. Specifically, the 3D data is converted to 2D attribute maps firstly. Then, the global estimation network is trained to predict facial landmarks roughly by feeding the fused CNN (Convolutional Neural Network) features extracted from facial attribute maps. After that, input the local fused CNN features extracted from the local patch around each landmark estimated previously, and other local models are trained separately to refine the locations. Tested on the Bosphorus and BU-3DFE datasets, the experimental results demonstrated effectiveness and accuracy of the proposed method for locating facial landmarks. Compared with existed methods, our results have achieved state-of-the-art performance.
Open AccessA Coarse-to-Fine Approach for 3D Facial Landmarking by Using Deep Feature Fusion
Facial landmarking locates the key facial feature points on facial data, which provides not only information on semantic facial structures, but also prior knowledge for other kinds of facial analysis. However, most of the existing works still focus on the 2D facial image which may suffer from lighting condition variations. In order to address this limitation, this paper presents a coarse-to-fine approach to accurately and automatically locate the facial landmarks by using deep feature fusion on 3D facial geometry data. Specifically, the 3D data is converted to 2D attribute maps firstly. Then, the global estimation network is trained to predict facial landmarks roughly by feeding the fused CNN (Convolutional Neural Network) features extracted from facial attribute maps. After that, input the local fused CNN features extracted from the local patch around each landmark estimated previously, and other local models are trained separately to refine the locations. Tested on the Bosphorus and BU-3DFE datasets, the experimental results demonstrated effectiveness and accuracy of the proposed method for locating facial landmarks. Compared with existed methods, our results have achieved state-of-the-art performance.
- 1 January 2006
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-5
Surface detection with subvoxel accuracy using facet model and IDDG operator
A surface detection algorithm with subvoxel accuracy is presented to extract edge and surface information from computed tomography (CT) images for virtual measurement and reverse engineering. The proposed algorithm employs the facet model and 3D IDDG operator to estimate the gradient vector and 3D directional derivatives. Subvoxel accuracy is achieved by locating the zeros of the 3D second directional derivative along the estimated gradient direction. Adaptive gradient threshold is used for the surface detector to improve the algorithm performance when applied to images with non-uniform contrast caused by CT artifacts, such as beam-hardening. Experiment results are presented to demonstrate the efficacy of the proposed method
- 1 January 2006
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-5
Surface detection with subvoxel accuracy using facet model and IDDG operator
A surface detection algorithm with subvoxel accuracy is presented to extract edge and surface information from computed tomography (CT) images for virtual measurement and reverse engineering. The proposed algorithm employs the facet model and 3D IDDG operator to estimate the gradient vector and 3D directional derivatives. Subvoxel accuracy is achieved by locating the zeros of the 3D second directional derivative along the estimated gradient direction. Adaptive gradient threshold is used for the surface detector to improve the algorithm performance when applied to images with non-uniform contrast caused by CT artifacts, such as beam-hardening. Experiment results are presented to demonstrate the efficacy of the proposed method