A5089128130- Cardiac Imaging and Diagnostics
- Face recognition and analysis
- Advanced X-ray and CT Imaging
- Face and Expression Recognition
- Medical Imaging Techniques and Applications
- Transportation Planning and Optimization
- Traffic Prediction and Management Techniques
- Advanced Image and Video Retrieval Techniques
- Human Mobility and Location-Based Analysis
- Human Pose and Action Recognition
- Video Analysis and Summarization
- Video Surveillance and Tracking Methods
- Image Retrieval and Classification Techniques
- Anomaly Detection Techniques and Applications
- Biometric Identification and Security
- Radiation Dose and Imaging
- Handwritten Text Recognition Techniques
- Advanced MRI Techniques and Applications
- Gait Recognition and Analysis
- Advanced Vision and Imaging
- Coronary Interventions and Diagnostics
- Optical Coherence Tomography Applications
- Astronomical Observations and Instrumentation
- Advanced Image Processing Techniques
- Generative Adversarial Networks and Image Synthesis
Karlsruhe Institute of Technology
2024
Urbana University
2024
Southwest Jiaotong University
2018-2023
Case Western Reserve University
2015-2020
LinkedIn (United States)
2020
University Hospitals Cleveland Medical Center
2016
Michigan State University
2015
Tsinghua University
2003-2014
University of Electronic Science and Technology of China
2013
Automatic face recognition is now widely used in applications ranging from deduplication of identity to authentication mobile payment. This popularity has raised concerns about spoof attacks (also known as biometric sensor presentation attacks), where a photo or video an authorized person's could be gain access facilities services. While number detection techniques have been proposed, their generalization ability not adequately addressed. We propose efficient and rather robust algorithm...
Short-term metro passenger flow forecasting is an essential component of intelligent transportation systems (ITS) and can be applied to optimize the organization a station offer data support for early warning system management. LSTM neural networks have recently achieved remarkable recent in field natural language processing (NLP) because they are well suited learning from experience predict time series. For this purpose, we propose empirical mode decomposition (EMD)-based long short-term...
Origin-destination (OD) short-term passenger flow forecasting (OD STPFF) in urban rail transit (URT) is essential for developing timely network measures. The capacity utilization ratios of critical sections are key factors these OD pairs passing through require a higher prediction accuracy than others; however, most studies have raised equal concerns on the each pair, namely, at level. To address this problem, we raise heterogeneous time-variant and employ an operation-oriented deep-learning...
When reading a document, glancing at the spatial layout of document is an initial step to understand it roughly. Traditional analysis (DLA) methods, however, offer only superficial parsing documents, focusing on basic instance detection and often failing capture nuanced logical relations between instances. These limitations hinder DLA-based models from achieving gradually deeper comprehension akin human reading. In this work, we propose novel graph-based Document Structure Analysis (gDSA)...
Congestion in urban rail transit (URT) systems often results passengers being left behind on platforms due to trains’ reaching capacity. Distinguishing between the travel choice behaviors of who board first arriving train (Type I passengers) and those are II passenger assignment is essential for effective URT management. This paper proposes a data‐driven passenger‐to‐train model (DPTAM) that leverages automated fare collection (AFC) data vehicle location (AVL) differentiate two types...
In recent years, image priors based on nonlocal self-similarity and low-rank approximation have been proven as powerful tools for restoration. Many restoration methods group similar patches a matrix recover the underlying structure from corrupted via rank minimization. However, both nonlocally redundant properties are highly content dependent, whether they can faithfully characterize wide range of natural images still remains unclear. this paper, we analyze these two provide quantifications...
People are often the most important subjects in photos, and ability of finding photos a particular person easily quickly an image collection is highly desired. In this paper, we present face clustering system which automatically groups into clusters, with each cluster containing same person. This done based on advanced recognition engine semi-supervised approach. The achieved good accuracy when tested different sets by users. Moreover, features such as adding new images, navigation retrieval...
High-quality Extragalactic Legacy-field Monitoring (HELM) is a long-term observing program that photometrically monitors several well-studied extragalactic legacy fields with the Dark Energy Camera (DECam) imager on CTIO 4m Blanco telescope. Since Feb 2019, HELM has been monitoring regions within COSMOS, XMM-LSS, CDF-S, S-CVZ, ELAIS-S1, and SDSS Stripe 82 few-day cadences in $(u)gri(z)$ bands, over collective sky area of $\sim 38$ deg${\rm ^2}$. The main science goal to provide high-quality...
High resolution, 100 frames/sec intravascular optical coherence tomography (IVOCT) can distinguish plaque types, but further validation is needed, especially for automated characterization. We developed experimental and 3D registration methods, to provide of IVOCT pullback volumes using microscopic, brightfield fluorescent cryo-image volumes, with optional, exactly registered cryo-histology. The innovation was a method match an pull-back images, acquired in the catheter reference frame, true...
Existing face recognition approaches are mostly developed based on adult faces which may not work well in distinguishing of kids. Especially, baby tend to have common features such as round cheeks and chins, so that current engines often fail differentiate them. In this paper, we present methods for discriminating from faces, training a special engine recognize different babies. To achieve these, collected huge number images software system annotate the image database. Experimental results...
In this paper, a new facial feature called Biologically Inspired Active Appearance Model (BIAAM) is proposed for face age estimation by using novel function learning algorithm, Local Ordinal Ranking (LOR). BIAAM, appearance variations are encoded extracting Bio Feature from normalized shape-free images with mean shape mask. The LOR divides the training set into several groups according to labels and applies Hyperplanes Ranker each group determine final predicting age. A multiple linear...
People are often the most important subjects in videos. It is highly desired to automatically summarize occurrences of different people a large collection video and quickly find clips containing particular person among them. In this paper, we present person-based summarization retrieval system named VideoWho which extracts temporal face sequences videos groups them into clusters, with each cluster same person. This accomplished based on advanced detection tracking algorithms, together...
We have demonstrated the ability to identify coronary calcium, a reliable biomarker of artery disease, using nongated, 2-shot, dual energy (DE) chest x-ray imaging. Here we will use digital simulations, backed up by measurements, characterize DE calcium signals and role potential confounds such as beam hardening, scatter, cardiac motion, pulmonary pulsation. For signal, consider quantification, compared CT score, visualization.We created stylized anatomical 3D phantoms including heart, lung,...
Evidence suggests high-resolution, high-contrast, [Formula: see text] intravascular optical coherence tomography (IVOCT) can distinguish plaque types, but further validation is needed, especially for automated characterization. We developed experimental and three-dimensional (3-D) registration methods to provide of IVOCT pullback volumes using microscopic, color, fluorescent cryo-image with optional registered cryo-histology. A specialized method matched images acquired in the catheter...
Age estimation via face images has recently attracted a lot of researches in computer vision, due to its many potential applications. In this paper, novel Bisection Search Tree (BST) algorithm is proposed for age estimation, based on the idea Divide and Conquer. Different from those conventional classification or regression approaches, BST first constructs binary tree according whole range training set, then learns decision functions all non-leaf nodes determine which child node test sample...
Coronary artery calcification (CAC) as assessed with CT calcium score is the best biomarker of coronary disease. Dual energy x-ray provides an inexpensive, low radiation-dose alternative. A two shot system (GE Revolution-XRd) used, raw images are processed a custom algorithm, and image (DECCI) created, similar to bone image, but optimized for CAC visualization, not lung visualization. In this report, we developed physicsbased, digital-phantom containing heart, lung, CAC, spine, ribs,...
Coronary artery calcification (CAC) as assessed with CT calcium score is the best biomarker for predicting future cardiac events. Dual energy (DE) chest radiography offers an inexpensive, low radiation-dose alternative to CT. We have shown CAC can be visualized using non-gated, 2-shot, DE x-ray imaging. However, signal from ribs, superimposed on signal, interfere both image registration and reader detection of CAC. To improve algorithm detectability CAC, we created enhanced visualization...
Image and video-based content has become ever present in a variety of domains like news, entertainment education. Users typically discover engage with via search recommendation systems. It is also important to serve high quality data users by filtering out irrelevant or harmful content. Thus, there an increasing need leverage the rich information image video order power systems for recommendation. At same time, effectiveness efficiency these been accelerated availability large-scale labeled...
The residual life prediction of aero-engine is important for ensuring flight safety and reducing operating costs airlines. Since there are varied performance parameters aero-engine, it difficult to use comprehensively these predict the life. This paper exploits Support Vector Regression Machine (SVR) in predicting trend aero-engine. Besides, a failure decision function based on Classification (SVC) established, which considered time wing. A method proposed function. approach applied data...
In this paper, we propose a novel generative method which could generate images with different illuminations by using single front-lighted sample. The is based on the linear Lambertian property and requires bootstrap set multiple subjects specific for each subject. During generation process, also scale decomposition to retain identity details between generated image original Unlike most of methods, proposed does NOT require 3D model. On other hand, proved flexible use because it be...
In this paper, a method for multi-sensor information fusion based on D-S evidence theory is discussed. An aero-engine gas path component used as the research object. The researches are executed fault level and decision level. effectiveness of proposed comprehensive diagnosis complex system illustrated by case study.