A5012943469- Face recognition and analysis
- Face and Expression Recognition
- Image Retrieval and Classification Techniques
- Advanced Image Processing Techniques
- Video Surveillance and Tracking Methods
- Advanced Image and Video Retrieval Techniques
- Biometric Identification and Security
- AI in cancer detection
- Functional Brain Connectivity Studies
- EEG and Brain-Computer Interfaces
- Vehicle License Plate Recognition
- IoT-based Smart Home Systems
- Image Processing Techniques and Applications
- Image and Signal Denoising Methods
- Domain Adaptation and Few-Shot Learning
- Logic, Reasoning, and Knowledge
- Retinal Imaging and Analysis
- Colorectal Cancer Screening and Detection
- Intelligent Tutoring Systems and Adaptive Learning
- Digital Imaging for Blood Diseases
- Digital Rights Management and Security
- Advanced Vision and Imaging
- Advanced Neural Network Applications
- Mental Health Research Topics
- AI-based Problem Solving and Planning
Center for Translational Research in Neuroimaging and Data Science
2024-2025
Georgia Institute of Technology
2024-2025
Emory University
2024-2025
University of Utah
2022
Gachon University
2017-2021
University of Tübingen
2014-2016
University of Wollongong
2006
Computer-aided diagnosis systems developed by computer vision researchers have helped doctors to recognize several endoscopic colorectal diseases more rapidly, which allows appropriate treatment and increases the patient's survival ratio. Herein, we present a robust architecture for image classification using an efficient dilation in Convolutional Neural Network (CNNs). It has high receptive field of view at deep layers increasing decreasing factor preserve spatial details. We argue that...
Colorectal cancer has become one of the most common cause mortality worldwide, with a five-year survival rate over 50%. Additionally, potential some polyp types to progress colorectal is considered high. Colonoscopy method for finding and removing polyps. However, during colonoscopy, significant number polyps missed as result human error mistakes. Thus, this study was primarily motivated by need obtain an early accurate diagnosis detected in colonoscopy images. In paper, we propose new...
Background: In functional magnetic resonance imaging (fMRI), network connectivity (FNC) captures temporal coupling among intrinsic networks (ICNs). Traditional FNC analyses often rely on linear models, which may overlook complex nonlinear interactions. We propose a multi-layered neural that generates heatmaps from matrices, we visualize at multiple layers, enabling us to better characterize multi-level interactions and improve interpretability. Methods: Our approach consists of two training...
Detecting and tracking humans are key problems for human-robot interaction. In this paper we present an algorithm mobile robots to detect track people reliably, even when go through different illumination conditions, often change in a wide variety of poses, frequently occluded. We have improved the performance face upper body detection quickly find each frame. This combination enhances efficiency human dealing with partial occlusions changes poses. To cope higher challenges complex poses...
Deep convolutional neural networks and generative adversarial currently attracted the attention of researchers because it is more effective than conventional representation-based methods. However, they have been facing two serious problems in trade-off between noise removal, artifacts, preserving low-contrast features high-frequency details. In particular, deep might fail to remove strong regions with higher levels while completely erasing By contrast, compared networks, be better balancing...
License plate image binarization is a critical step in Automatic Number Plate Recognition(ANPR) systems and essential for character segmentation. Generally Otsu(global) or adaptive(local) thresholding methods are commonly used, but each of them may have shortcoming terms segmenting all the characters accurately Optical Character Recognition(OCR) reading when not cropped exactly. In this paper we propose feedback based approach which fuses global local methods. Local method applied first to...
In functional magnetic resonance imaging (fMRI) studies, it is common to evaluate the brain's network connectivity (FNC) which captures temporal coupling between hemodynamic signals. FNC has been linked various psychological phenomena. However, current FNCs mainly represent linear statistical relationships, may not capture fully complexity of interactions among brain intrinsic networks (ICNs). Therefore, crucial explore approaches that can better account for possible intricate nonlinear...
One of the key challenges in use resting brain functional magnetic resonance imaging (fMRI) network analysis for predicting mental illnesses such as schizophrenia (SZ) is high noise levels variability among individuals including age, sex, and different protocols used labs. To deal with these challenging problems, we designed a recognition method using networks to classify SZs healthy controls (HCs). Our includes two stages training. In first stage, deep convolutional neural (DCNN) extract...