A5100716029- Face recognition and analysis
- Video Surveillance and Tracking Methods
- Face and Expression Recognition
- Functional Brain Connectivity Studies
- Human Pose and Action Recognition
- Image Enhancement Techniques
- Advanced Image Fusion Techniques
- EEG and Brain-Computer Interfaces
- Advanced Image and Video Retrieval Techniques
- Biometric Identification and Security
- Advanced Neural Network Applications
- Vehicle License Plate Recognition
- Medical Image Segmentation Techniques
- Advanced Neuroimaging Techniques and Applications
- Neural dynamics and brain function
- Advanced Image Processing Techniques
- Advanced MRI Techniques and Applications
- Generative Adversarial Networks and Image Synthesis
- Anomaly Detection Techniques and Applications
- Interconnection Networks and Systems
- Remote Sensing and LiDAR Applications
- Emotion and Mood Recognition
- Industrial Vision Systems and Defect Detection
- Advanced Vision and Imaging
- Image and Signal Denoising Methods
Jeonbuk National University
2014-2023
Hitachi Global Storage Technologies (United States)
2023
University Medical Center Freiburg
2022
University of Freiburg
2022
Montreal Neurological Institute and Hospital
2022
McGill University
2022
Chonbuk National University Hospital
2004-2017
BRAC University
2013
Korea Advanced Institute of Science and Technology
2007
Analyses of gray matter concentration (GMC) deficits in patients with schizophrenia (Sz) have identified robust changes throughout the cortex. We assessed relationships between diagnosis, overall symptom severity, and patterns largest aggregated structural imaging dataset to date. performed both source-based morphometry (SBM) voxel-based (VBM) analyses on GMC images from 784 Sz 936 controls (Ct) across 23 scanning sites Europe United States. After correcting for age, gender, site, diagnosis...
Recent advancements in human-computer interaction research have led to the possibility of emotional communication via brain-computer interface systems for patients with neuropsychiatric disorders or disabilities. In this paper, we efficiently recognize states by analyzing features electroencephalography (EEG) signals, which are generated from EEG sensors that noninvasively measure electrical activity neurons inside human brain, and select optimal combination these recognition. scalp data 21...
Make and model recognition (MMR) of vehicles plays an important role in automatic vision-based systems. This paper proposes a novel deep learning approach for MMR using the SqueezeNet architecture. The frontal views vehicle images are first extracted fed into network training testing. architecture with bypass connections between Fire modules, variant vanilla SqueezeNet, is employed this study, which makes our system more efficient. experimental results on collected large-scale datasets...
Schizophrenia is a complex, debilitating mental disorder characterized by wide-ranging symptoms including delusions, hallucinations (so-called positive symptoms), and impaired motor speech/language production negative symptoms). Salience-monitoring theorists propose that abnormal functional communication between the salience network (SN) default mode (DMN) begets of schizophrenia, yet prior studies have predominately reported links disrupted SN/DMN symptoms. It remains unclear whether...
Human action monitoring can be advantageous to remotely monitor the status of patients or elderly person for intelligent healthcare. recognition enables efficient and accurate human behaviors, which exhibit multifaceted complexity attributed disparities in viewpoints, personality, resolution motion speed individuals, etc. The spatial-temporal information plays an important role recognition. In this paper, we proposed a novel deep learning architecture named as recurrent 3D convolutional...
Emotion recognition is an important task for computer to understand the human status in brain interface (BCI) systems. It difficult perceive emotion of some disabled people through their facial expression, such as functional autism patient. EEG signal provides us a non-invasive way recognize these disable headset electrodes placed on scalp. In this paper, we propose deep learning algorithm simultaneously learn features and classify emotions signals. differs from conventional methods apply...
Affective computing research field is growing in large scale with the frequent development of human-computer applications. These applications use information mental or affective conditions desired subjects to train their brain responses. Facial impressions, text physiology, vocal and other about are used classification algorithms. However, frameworks for EEG signals have been infrequently due lack a complete theoretical framework. Therefore, we present here an analysis two different methods...
Vehicle analysis involves license-plate recognition (LPR), vehicle-type classification (VTC), and vehicle make model (MMR). Among these tasks, MMR plays an important complementary role in respect to LPR. In this paper, we propose a novel framework for using local tiled deep networks. The frontal views of images are first extracted fed into the networks training testing. A convolutional neural network (LTCNN) is proposed alter weight sharing scheme CNN with structure. LTCNN unties weights...
Prior resting-state functional magnetic resonance imaging (fMRI) analyses have identified patterns of connectivity associated with hallucinations in schizophrenia (Sz). In this study, we performed an analysis the mean amplitude low-frequency fluctuations (ALFF) to compare resting state spontaneous patients Sz who report experiencing impacting different sensory modalities. By exploring dynamics across 2 passbands (slow-4 and slow-5), assessed impact hallucination modality frequency range on...
Human computer interaction is a growing field in terms of helping people their daily life to improve living. Especially, with some disability may need an interface which more appropriate and compatible needs. Our research focused on similar kinds problems, such as students mental disorder or mood disruption problems. To learning process, intelligent emotion recognition system essential has ability recognize the current emotional state brain. Nowadays, special schools, instructors are...
Natural environments usually have a larger dynamic range than the that can be acquired by an optical camera with single shot. In this paper, we propose multiexposure fusion method effectively fuses in direct manner differently exposed images of high scene into high-quality image. First, present developed joint weight considering exposure level measurement local and global luminance components input images. Second, introduce designed multiscale edge-preserving smoothing (MEPS) model for...
We propose and develop a novel biclustering (N-BiC) approach for performing N-way of neuroimaging data. Our is applicable to an arbitrary number features from both imaging behavioral data (e.g., symptoms). applied it structural MRI patients with schizophrenia.It uses source-based morphometry [i.e., independent component analysis gray matter segmentation maps] decompose the into set spatial maps, each which includes regions that covary among individuals. Then, loading parameters components...
Human pose estimation is a problem that continues to be one of the greatest challenges in field computer vision. While stacked structure an hourglass network has enabled substantial progress human and key-point detection areas, it largely used as backbone network. However, also requires relatively large number parameters high computational capacity due characteristics its structure. Accordingly, present work proposes more lightweight version network, which improves performance. The new...
Clinical and cognitive symptoms domain-based subtyping in schizophrenia (Sz) has been critiqued due to the lack of neurobiological correlates heterogeneity symptom scores. We, therefore, present a novel data-driven framework using biclustered independent component analysis detect subtypes from reliable stable gray matter concentration (GMC) patients with Sz. The developed methodology consists following steps: source-based morphometry (SBM) decomposition, selection sorting two loadings,...
In rainy conditions, imaging devices often capture degraded and blurry images. Most existing works on this problem focus rain streak removal, but these approaches cannot handle the various types of found in paper, we propose a robust removal method for use with single images using an attentive composite residual network. We put forth single-to-dual encoder-decoder structure, which consists net that identifies regions containing components during encoding, followed by dual-channel...
Skeleton-based action recognition is a widely used task in related research because of its clear features and the invariance human appearances illumination. Furthermore, it can also effectively improve robustness recognition. Graph convolutional networks have been implemented on those skeletal data to recognize actions. Recent studies shown that graph neural network works well using spatial temporal skeleton data. The prevalent methods extract purely rely deep learn from primitive 3D...
Face sketch synthesis aims to generate a face image from corresponding photo and has wide applications in law enforcement digital entertainment. Despite the remarkable achievements that have been made synthesis, most existing works pay main attention facial content transfer, at expense of detail information. In this paper, we present new generative adversarial learning framework focus on preservation for realistic synthesis. Specifically, high-resolution network is modified as generator...
Vehicle analysis has been investigated for decades, which involves license plate recognition, intelligent traffic. Among these applications, vehicle make recognition is a challenging task due to the close appearance between car models. In this paper, we propose an architecture recognize based on convolutional neural network (CNN). The moving first localized by frame difference, resultant binary image used detect frontal view of symmetry filter. detected train and test CNN. Experimental...
Person re-identification (re-ID) has been gaining in popularity the research community owing to its numerous applications and growing importance surveillance industry. Recent methods often employ partial features for person re-ID offer fine-grained information beneficial retrieval. In this paper, we focus on learning improved discriminative using a deep convolutional neural architecture, which includes pyramid spatial pooling module efficient feature representation. Furthermore, propose...