A5053808377- EEG and Brain-Computer Interfaces
- Functional Brain Connectivity Studies
- Neural dynamics and brain function
- Brain Tumor Detection and Classification
- Advanced Image Processing Techniques
- Fault Detection and Control Systems
- Image and Video Stabilization
- Blind Source Separation Techniques
- Text and Document Classification Technologies
- Web Data Mining and Analysis
- Algorithms and Data Compression
- Image Processing Techniques and Applications
- Sparse and Compressive Sensing Techniques
- Speech and Audio Processing
- Target Tracking and Data Fusion in Sensor Networks
- Advanced MRI Techniques and Applications
Stevens Institute of Technology
2023-2024
Software (Spain)
2021
Peking University
2003-2021
Abstract EEG/MEG source imaging (ESI) aims to find the underlying brain sources explain observed EEG or MEG measurement. Multiple classical approaches have been proposed solve ESI problem based on different neurophysiological assumptions. To support clinical decision-making, it is important estimate not only exact location of signal but also extended activation regions. Existing methods may render over-diffuse sparse solutions, which limit extent estimation accuracy. In this work, we...
The process of reconstructing underlying cortical and subcortical electrical activities from Electroencephalography (EEG) or Magnetoencephalography (MEG) recordings is called Electrophysiological Source Imaging (ESI). Given the complementarity between EEG MEG in measuring radial tangential sources, combined EEG/MEG considered beneficial improving reconstruction performance ESI algorithms. Traditional algorithms mainly emphasize incorporating predesigned neurophysiological priors to solve...
Automatic feature selection methods such as document frequency, information gain, mutual and so on are commonly applied in the preprocess of text categorization order to reduce originally high dimension a bearable level, meanwhile also noise improve precision. Generally they assess specific term by calculating its occurrences among individual categories or entire corpus, where "occurring document" is simply defined occurring at least once. A major drawback this measure that, for single...
In face recognition systems, highly robust facial feature representation and good classification algorithm performance can affect the effect of under unrestricted conditions. To explore anti‐interference convolutional neural network (CNN) reconstructed by deep learning (DL) framework in image extraction (FE) recognition, paper, first, inception structure GoogleNet residual error ResNet are combined to construct a new reconstruction algorithm, with random gradient descent (SGD) triplet loss...
The consumption of high doses marijuana can have significant psychological and social impacts. In this study, we propose an interpretable novel framework called the HOGAB (High-Order Graph Attention Neural Networks) model for addictive Marijuana classification analysis localized network clusters that demonstrated abnormal brain activities among chronic users. integrates dynamic intrinsic functional networks with LSTM technology to capture temporal patterns in fMRI time series We employed...
Abstract Activity in the human brain is composed of complex firing patterns and interactions among neurons neuronal circuits. The neuroimaging field underwent a paradigm shift over past decades from mapping tasked evoked regions activations towards identifying characterizing dynamic networks coordinating regions. Electrophysiological signals are direct manifestation activities, thus whole electrophysiological (WBEN) can serve as fundamental tool for neuroscience studies clinical...
Abstract High Frequency Oscillations (HFOs) is an important biomarker that can potentially pinpoint the epileptogenic zones (EZs). However, duration of HFO short with around 4 cycles, which might be hard to recognize when embedded within signals lower frequency oscillatory background. In addition, annotating HFOs manually time-consuming given long-time recordings and up hundreds intracranial electrodes. We propose leverage a Switching State Space Model (SSSM) identify events automatically...