A5028418236- Radiomics and Machine Learning in Medical Imaging
- AI in cancer detection
- MRI in cancer diagnosis
- Digital Radiography and Breast Imaging
- COVID-19 diagnosis using AI
- Human Pose and Action Recognition
- Global Cancer Incidence and Screening
- Medical Imaging Techniques and Applications
- Artificial Intelligence in Healthcare and Education
- Traumatic Brain Injury and Neurovascular Disturbances
- Privacy-Preserving Technologies in Data
- Colorectal Cancer Screening and Detection
- Video Surveillance and Tracking Methods
- Gastric Cancer Management and Outcomes
- Advanced MRI Techniques and Applications
- Advanced Vision and Imaging
- Advanced Image and Video Retrieval Techniques
- Anomaly Detection Techniques and Applications
- Advanced X-ray and CT Imaging
- Breast Cancer Treatment Studies
- Advanced Neuroimaging Techniques and Applications
- Traumatic Brain Injury Research
- Imbalanced Data Classification Techniques
- Medical Imaging and Analysis
- BRCA gene mutations in cancer
University of Pittsburgh
2016-2025
University of Pittsburgh Medical Center
2024
UPMC Health System
2023
Craft Engineering Associates (United States)
2020-2022
Intelligent Systems Research (United States)
2022
Ninghai First Hospital Medicare and Health Group
2021
Oxford University Press (United Kingdom)
2020
University of Alberta
2020
Hospital of the University of Pennsylvania
2015
University of Pennsylvania
2012-2013
A novel method for crowd flow modeling and anomaly detection is proposed both coherent incoherent scenes. The novelty revealed in three aspects. First, it a unique utilization of particle trajectories crowded scenes, which we propose new efficient representative arbitrarily complicated flows. Second, chaotic dynamics are introduced into the context to characterize motions by regulating set invariant features, reliably computed used detecting anomalies. Third, probabilistic framework...
Mammographic breast density is an established risk marker for cancer and visually assessed by radiologists in routine mammogram image reading, using four qualitative Breast Imaging Reporting Data System (BI-RADS) categories. It particularly difficult to consistently distinguish the two most common variably assigned BI-RADS categories, i.e., "scattered density" "heterogeneously dense". The aim of this work was investigate a deep learning-based classifier these aiming at providing potential...
Recognition of human actions in a video acquired by moving camera typically requires standard preprocessing steps such as motion compensation, object detection and tracking. The errors from the compensation step propagate to stage, resulting miss-detections, which further complicates tracking cluttered incorrect tracks. Therefore, action recognition is considered very challenging. In this paper, we propose novel approach does not follow steps, accordingly avoids aforementioned difficulties....
Abstract Purpose: False positives in digital mammography screening lead to high recall rates, resulting unnecessary medical procedures patients and health care costs. This study aimed investigate the revolutionary deep learning methods distinguish recalled but benign images from negative exams those with malignancy. Experimental Design: Deep convolutional neural network (CNN) models were constructed classify into malignant (breast cancer), cancer free), recalled-benign categories. A total of...
Purpose To investigate two deep learning‐based modeling schemes for predicting short‐term risk of developing breast cancer using prior normal screening digital mammograms in a case‐control setting. Methods We conducted retrospective Institutional Review Board‐approved study on cohort 226 patients (including 113 women diagnosed with and controls) who underwent general population screening. For each patient, (i.e., negative or benign findings) mammogram examination [including mediolateral...
Background After severe traumatic brain injury (sTBI), physicians use long-term prognostication to guide acute clinical care yet struggle predict outcomes in comatose patients. Purpose To develop and evaluate a prognostic model combining deep learning of head CT scans information after sTBI. Materials Methods This was retrospective analysis two prospectively collected databases. The model-building set included 537 patients (mean age, 40 years ± 17 [SD]; 422 men) from one institution November...
Conventional federated learning (FL) trains one global model for a federation of clients with decentralized data, reducing the privacy risk centralized training. However, distribution shift across non-IID datasets, often poses challenge to this one-model-fits-all solution. Personalized FL aims mitigate issue systematically. In work, we propose APPLE, personalized cross-silo framework that adaptively learns how much each client can benefit from other clients' models. We also introduce method...
Purpose: Breast magnetic resonance imaging (MRI) plays an important role in the clinical management of breast cancer. Studies suggest that relative amount fibroglandular (i.e., dense) tissue as quantified MR images can be predictive risk for developing cancer, especially high‐risk women. Automated segmentation and volumetric density estimation MRI could therefore useful cancer assessment. Methods: In this work authors develop validate a fully automated algorithm, namely, atlas‐aided fuzzy...
Purpose: Breast magnetic resonance imaging (MRI) plays an important role in the clinical management of breast cancer. Computerized analysis is increasingly used to quantify MRI features applications such as computer‐aided lesion detection and fibroglandular tissue estimation for cancer risk assessment. Automated segmentation whole‐breast organ from other parts imaged step aiding localization quantification. For this task, identifying chest wall line (CWL) most challenging due image contrast...
Background The axillary lymph node status is critical for breast cancer staging and individualized treatment planning. Purpose To assess the effect of determining (ALN) metastasis by MRI‐derived radiomic signatures, compare discriminating abilities different MR sequences. Study Type Retrospective. Population In all, 120 patients, 59 with ALN 61 without metastasis, all confirmed pathology. Field Strength/Sequence 3 .0T scanner T 1 ‐weighted imaging, 2 diffusion‐weighted dynamic...
Adrenaline and noradrenaline are produced within the heart from neuronal non-neuronal sources. These adrenergic hormones have profound effects on cardiovascular development function, yet relatively little information is available about specific tissue distribution of cells adult heart. The purpose present study was to define anatomical localization derived an lineage To accomplish this, we performed genetic fate-mapping experiments where mice with cre-recombinase (Cre) gene inserted into...
We present a fully automated method for deriving quantitative measures of background parenchymal enhancement (BPE) from breast dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) and perform preliminary evaluation these to assess the effect risk-reducing salpingo-oophorectomy (RRSO) in cohort cancer susceptibility gene 1/2 (BRCA1/2) mutation carriers.Breast DCE-MRI data 50 BRCA1/2 carriers were retrospectively analyzed compliance with Health Insurance Portability Accountability...
The purpose of this study was to distinguish axillary lymph node (ALN) status using preoperative breast DCE-MRI radiomics and compare the effects two-dimensional (2D) three-dimensional (3D) analysis.A retrospective including 154 cancer patients all confirmed by pathology; 80 with ALN metastasis 74 without. All MRI scans were achieved at a 3.0 Tesla scanner 7 post-contrast MR phases sequentially acquired temporal resolution 60 s. radiomic features extracted separately from 2D single slice...
Background Diffusion‐weighted imaging (DWI) in MRI plays an increasingly important role diagnostic applications and developing biomarkers. Automated whole‐breast segmentation is yet challenging step for quantitative breast analysis. While methods have been developed on dynamic contrast‐enhanced (DCE) MRI, automatic DWI still underdeveloped. Purpose To develop a deep/transfer learning‐based approach scans conduct extensive study assessment four datasets from both internal external sources....
Abstract While active efforts are advancing medical artificial intelligence (AI) model development and clinical translation, safety issues of the AI models emerge, but little research has been done. We perform a study to investigate behaviors an diagnosis under adversarial images generated by Generative Adversarial Network (GAN) evaluate effects on human experts when visually identifying potential images. Our GAN makes intentional modifications diagnosis-sensitive contents mammogram in deep...
Background parenchymal enhancement (BPE) at dynamic contrast-enhanced (DCE) MRI of cancer-free breasts increases the risk developing breast cancer; implications quantitative BPE in ipsilateral with cancer are largely unexplored. Purpose To determine whether measurements one or both could be used to predict recurrence women cancer, using Oncotype DX score as reference standard. Materials and Methods This HIPAA-compliant retrospective single-institution study included diagnosed between January...