A5043710204- Radiomics and Machine Learning in Medical Imaging
- Advanced Neural Network Applications
- COVID-19 diagnosis using AI
- Medical Image Segmentation Techniques
- AI in cancer detection
- Privacy-Preserving Technologies in Data
- Medical Imaging and Analysis
- Colorectal Cancer Screening and Detection
- Medical Imaging Techniques and Applications
- Advanced X-ray and CT Imaging
- Domain Adaptation and Few-Shot Learning
- Artificial Intelligence in Healthcare and Education
- Lung Cancer Diagnosis and Treatment
- Prostate Cancer Diagnosis and Treatment
- Colorectal Cancer Surgical Treatments
- Advanced Image and Video Retrieval Techniques
- Pancreatic and Hepatic Oncology Research
- Adversarial Robustness in Machine Learning
- Machine Learning and Data Classification
- Advanced MRI Techniques and Applications
- Integrated Circuits and Semiconductor Failure Analysis
- Retinal Imaging and Analysis
- Advancements in Photolithography Techniques
- Cardiac Imaging and Diagnostics
- Brain Tumor Detection and Classification
Nvidia (United States)
2019-2025
Baker Hughes (Germany)
2023
Santa Clara University
2021-2022
Nvidia (United Kingdom)
2018-2022
Stanford University
2022
Nagoya University
2015-2021
National Institutes of Health Clinical Center
2014-2019
Mass General Brigham
2019
National Institutes of Health
2015-2017
University of Tsukuba
2016
Remarkable progress has been made in image recognition, primarily due to the availability of large-scale annotated datasets and deep convolutional neural networks (CNNs). CNNs enable learning data-driven, highly representative, hierarchical features from sufficient training data. However, obtaining as comprehensively ImageNet medical imaging domain remains a challenge. There are currently three major techniques that successfully employ classification: CNN scratch, using off-the-shelf...
Fully Convolutional Neural Networks (FCNNs) with contracting and expanding paths have shown prominence for the majority of medical image segmentation applications since past decade. In FCNNs, encoder plays an integral role by learning both global local features contextual representations which can be utilized semantic output prediction decoder. Despite their success, locality convolutional layers in limits capability long-range spatial dependencies. Inspired recent success transformers...
Automated computer-aided detection (CADe) in medical imaging has been an important tool clinical practice and research. State-of-the-art methods often show high sensitivities but at the cost of false-positives (FP) per patient rates. We design a two-tiered coarse-to-fine cascade framework that first operates candidate generation system $\sim$100% FP levels. By leveraging existing CAD systems, coordinates regions or volumes interest (ROI VOI) for lesion candidates are generated this step...
Chest CT is emerging as a valuable diagnostic tool for clinical management of COVID-19 associated lung disease. Artificial intelligence (AI) has the potential to aid in rapid evaluation scans differentiation findings from other entities. Here we show that series deep learning algorithms, trained diverse multinational cohort 1280 patients localize parietal pleura/lung parenchyma followed by classification pneumonia, can achieve up 90.8% accuracy, with 84% sensitivity and 93% specificity,...
Federated learning (FL) is a method used for training artificial intelligence models with data from multiple sources while maintaining anonymity, thus removing many barriers to sharing. Here we 20 institutes across the globe train FL model, called EXAM (electronic medical record (EMR) chest X-ray AI model), that predicts future oxygen requirements of symptomatic patients COVID-19 using inputs vital signs, laboratory and X-rays. achieved an average area under curve (AUC) >0.92 predicting...
Vision Transformers (ViT)s have shown great performance in self-supervised learning of global and local representations that can be transferred to downstream applications. Inspired by these results, we introduce a novel framework with tailored proxy tasks for medical image analysis. Specifically, propose: (i) new 3D transformer-based model, dubbed Swin UNEt TRansformers (Swin UNETR), hierarchical encoder pretraining; (ii) the underlying pattern human anatomy. We demonstrate successful...
Recent advances in deep learning for medical image segmentation demonstrate expert-level accuracy. However, application of these models clinically realistic environments can result poor generalization and decreased accuracy, mainly due to the domain shift across different hospitals, scanner vendors, imaging protocols, patient populations etc. Common transfer adaptation techniques are proposed address this bottleneck. solutions require data (and annotations) from target retrain model, is...
The adenoma detection rate is an established quality indicator for colonoscopy. For instance, a 1% increase in the was associated with 3% decrease interval colorectal cancer incidence.1Corley D.A. Jensen C.D. Marks A.R. et al.Adenoma and risk of death.N Engl J Med. 2014; 370: 1298-1306Crossref PubMed Scopus (1166) Google Scholar However, previous meta-analysis showed that approximately 26% neoplastic diminutive polyps were missed single colonoscopy.2van Rijn J.C. Reitsma J.B. Stoker J....
Interstitial lung diseases (ILD) involve several abnormal imaging patterns observed in computed tomography (CT) images. Accurate classification of these plays a significant role precise clinical decision making the extent and nature diseases. Therefore, it is important for developing automated pulmonary computer-aided detection systems. Conventionally, this task relies on experts' manual identification regions interest (ROIs) as prerequisite to diagnose potential This protocol time consuming...
Artificial Intelligence (AI) is having a tremendous impact across most areas of science. Applications AI in healthcare have the potential to improve our ability detect, diagnose, prognose, and intervene on human disease. For models be used clinically, they need made safe, reproducible robust, underlying software framework must aware particularities (e.g. geometry, physiology, physics) medical data being processed. This work introduces MONAI, freely available, community-supported,...
Robust organ segmentation is a prerequisite for computer-aided diagnosis, quantitative imaging analysis, pathology detection, and surgical assistance. For organs with high anatomical variability (e.g., the pancreas), previous approaches report low accuracies, compared well-studied organs, such as liver or heart. We present an automated bottom-up approach pancreas in abdominal computed tomography (CT) scans. The method generates hierarchical cascade of information propagation by classifying...
The diagnostic performance of CT for pancreatic cancer is interpreter-dependent, and approximately 40% tumours smaller than 2 cm evade detection. Convolutional neural networks (CNNs) have shown promise in image analysis, but the networks' potential detection diagnosis unclear. We aimed to investigate whether CNN could distinguish individuals with without on CT, compared radiologist interpretation.In this retrospective, study, contrast-enhanced images 370 patients 320 controls from a...
Abstract Objective To demonstrate enabling multi-institutional training without centralizing or sharing the underlying physical data via federated learning (FL). Materials and Methods Deep models were trained at each participating institution using local clinical data, an additional model was FL across all of institutions. Results We found that exhibited superior performance generalizability to single institutions, with overall level significantly better than any institutional alone when...
While making a tremendous impact in various fields, deep neural networks usually require large amounts of labeled data for training which are expensive to collect many applications, especially the medical domain. Un-labeled data, on other hand, is much more abundant. Semi-supervised learning techniques, such as co-training, could provide powerful tool leverage unlabeled data. In this paper, we propose novel framework, uncertainty-aware multi-view co-training (UMCT), address semi-supervised...
3D convolution neural networks (CNN) have been proved very successful in parsing organs or tumours medical images, but it remains sophisticated and time-consuming to choose design proper given different task contexts. Recently, Neural Architecture Search (NAS) is proposed solve this problem by searching for the best network architecture automatically. However, inconsistency between search stage deployment often exists NAS algorithms due memory constraints large space, which could become more...
Background Approximately 40% of pancreatic tumors smaller than 2 cm are missed at abdominal CT. Purpose To develop and to validate a deep learning (DL)-based tool able detect cancer Materials Methods Retrospectively collected contrast-enhanced CT studies in patients diagnosed with between January 2006 July 2018 were compared individuals normal pancreas (control group) obtained 2004 December 2019. An end-to-end comprising segmentation convolutional neural network (CNN) classifier ensembling...