A5076952204- Medical Image Segmentation Techniques
- Advanced MRI Techniques and Applications
- Advanced Neural Network Applications
- Advanced Image and Video Retrieval Techniques
- Medical Imaging Techniques and Applications
- Network Security and Intrusion Detection
- Robotics and Sensor-Based Localization
- Radiomics and Machine Learning in Medical Imaging
- Domain Adaptation and Few-Shot Learning
- Internet Traffic Analysis and Secure E-voting
- Retinal Imaging and Analysis
- Image Retrieval and Classification Techniques
- Lung Cancer Diagnosis and Treatment
- Glaucoma and retinal disorders
- COVID-19 diagnosis using AI
- Fetal and Pediatric Neurological Disorders
- Traumatic Brain Injury and Neurovascular Disturbances
- Spam and Phishing Detection
- Video Surveillance and Tracking Methods
- Advanced Malware Detection Techniques
- Acute Ischemic Stroke Management
- Advanced Bandit Algorithms Research
- Retinal and Optic Conditions
- Reinforcement Learning in Robotics
- Cerebrovascular and Carotid Artery Diseases
Cisco Systems (Czechia)
2015-2017
Czech Technical University in Prague
2016
Siemens (Germany)
2009-2014
Siemens (United States)
2011-2014
Princeton University
2010-2013
Rensselaer Polytechnic Institute
2006-2010
National Center for Nanoscience and Technology
2007
Motivated by the goals of improving detection low-contrast and narrow vessels eliminating false detections at nonvascular structures, a new technique is presented for extracting in retinal images. The core likelihood ratio test that combines matched-filter responses, confidence measures vessel boundary measures. Matched filter responses are derived scale-space to extract widely varying widths. A measure defined as projection vector formed from normalized pixel neighborhood onto ideal...
Multi-contrast MRI (MC-MRI) captures multiple complementary imaging modalities to aid in radiological decision-making. Given the need for lowering time cost of acquisitions, current deep accelerated reconstruction networks focus on exploiting redundancy between contrasts. However, existing works are largely supervised with paired data and/or prohibitively expensive fully-sampled sequences. Further, typically rely convolutional architectures which limited their capacity model long-range...
Our goal is an automated 2D-image-pair registration algorithm capable of aligning images taken a wide variety natural and man-made scenes as well many medical images. The should handle low overlap, substantial orientation scale differences, large illumination variations, physical changes in the scene. An important component this ability to automatically reject pairs that have no overlap or too differences be aligned well. We propose complete including techniques for initialization,...
Abstract Neuroimaging is crucial for assessing mass effect in brain-injured patients. Transport to an imaging suite, however, challenging critically ill We evaluated the use of a low magnetic field, portable MRI (pMRI) midline shift (MLS). In this observational study, 0.064 T pMRI exams were performed on stroke patients admitted neuroscience intensive care unit at Yale New Haven Hospital. Dichotomous (present or absent) and continuous MLS measurements obtained locally available accessible...
In this paper, we investigate the effect of substantial inter-image intensity changes and in modality on performance keypoint detection, description, matching algorithms context image registration. doing so, modify widely-used descriptors such as SIFT shape contexts, attempting to capture insight that some structural information is indeed preserved between images despite dramatic appearance changes. These extensions include (a) pairing opposite-direction gradients formation orientation...
Routine ultrasound exam in the second and third trimesters of pregnancy involves manually measuring fetal head brain structures 2-D scans. The procedure requires a sonographer to find standardized visualization planes with probe place measurement calipers on interest. process is tedious, time consuming, introduces user variability into measurements. This paper proposes an automatic (AFHB) system for automatically anatomical from 3-D volumes. searches volume hierarchy resolutions by focusing...
This paper proposes a new registration algorithm, Co-variance Driven Correspondences (CDC), that depends fundamentally on the estimation of uncertainty in point correspondences. is derived from covariance matrices individual locations and matrix estimated transformation parameters. Based this uncertainty, CDC uses robust objective function an EM-like algorithm to simultaneously estimate parameters, their matrix, likely Unlike Robust Point Matching (RPM) requires neither annealing schedule...
Retinal clinicians and researchers make extensive use of images, the current emphasis is on digital imaging retinal fundus. The goal this paper to introduce a system, known as image vessel extraction registration which provides community clinicians, researchers, study directors an integrated suite advanced analysis tools over Internet. capabilities include vasculature tracing morphometry, joint (simultaneous) montaging multiple fields, cross-modality (color/red-free fundus photographs...
In this paper, we propose a novel framework for detecting multiple objects in 2D and 3D images. Since joint multi-object model is difficult to obtain most practical situations, focus here on the sequentially, one-by-one. The interdependence of object poses strong prior information embedded our domain medical images results better performance than individually. Our approach based Sequential Estimation techniques, frequently applied visual tracking. Unlike tracking, where sequential order...
Motivated by the goals of automatically extracting vessel segments and constructing retinal vascular trees with anatomical realism, this paper presents analyses an algorithm that combines segmentation grouping extracted segments. The proposed method aims to restore topology realism for clinical studies diagnosis diseases, which manifest abnormalities in either venous and/or arterial systems. Vessel are grouped using extended Kalman filter takes into account continuities curvature, width,...
The expanding role of complex object detection algorithms introduces a need for flexible architectures that simplify interfacing with machine learning techniques and offer easy-to-use training procedures. To address this need, the Integrated Detection Network (IDN) proposes conceptual design rapid prototyping boundary systems. IDN uses strong spatial prior present in medical imaging domain large annotated database images to train robust detectors. best hypotheses are propagated throughout...