A5043572938- Advanced MRI Techniques and Applications
- Advanced Neuroimaging Techniques and Applications
- Atomic and Subatomic Physics Research
- MRI in cancer diagnosis
- Medical Imaging Techniques and Applications
- EEG and Brain-Computer Interfaces
- Photoacoustic and Ultrasonic Imaging
- Machine Learning in Healthcare
- Brain Tumor Detection and Classification
- Neural Networks and Applications
- Medical Image Segmentation Techniques
- Functional Brain Connectivity Studies
- Scientific Computing and Data Management
- Nuclear Physics and Applications
Massachusetts General Hospital
2024
Harvard University
2024
University of North Carolina at Chapel Hill
2021
Data-driven machine learning has made significant strides in medical image analysis. However, most existing methods are tailored to specific modalities and assume a particular resolution (often isotropic). This limits their generalizability clinical settings, where variations scan appearance arise from differences sequence parameters, resolution, orientation. Furthermore, general-purpose models designed for healthy subjects suffer performance degradation when pathology is present. We...
Brain atrophy and white matter hyperintensity (WMH) are critical neuroimaging features for ascertaining brain injury in cerebrovascular disease multiple sclerosis. Automated segmentation quantification is desirable but existing methods require high-resolution MRI with good signal-to-noise ratio (SNR). This precludes application to clinical low-field portable (pMRI) scans, thus hampering large-scale tracking of WMH progression, especially underserved areas where pMRI has huge potential. Here...
Perfusion imaging is of great clinical importance and used to assess a wide range diseases including strokes brain tumors. Commonly approaches for the quantitative analysis perfusion images are based on measuring effect contrast agent moving through blood vessels into tissue. Contrast-agent free approaches, example, intravoxel incoherent motion arterial spin labeling, also exist, but so far not routinely clinically. Existing contrast-agent-dependent methods typically rely estimation input...
Recent learning-based approaches have made astonishing advances in calibrated medical imaging like computerized tomography (CT), yet they struggle to generalize uncalibrated modalities -- notably magnetic resonance (MR) imaging, where performance is highly sensitive the differences MR contrast, resolution, and orientation. This prevents broad applicability diverse real-world clinical protocols. We introduce Brain-ID, an anatomical representation learning model for brain imaging. With...
Remarkable progress has been made by data-driven machine-learning methods in the analysis of MRI scans. However, most existing approaches are crafted for specific MR pulse sequences (MR contrasts) and usually require nearly isotropic acquisitions. This limits their applicability to diverse real-world clinical data, where scans commonly exhibit variations appearances due being obtained with varying sequence parameters, resolutions, orientations -- especially presence pathology. In this paper,...
Over recent years, deep learning based image registration has achieved impressive accuracy in many domains, including medical imaging and, specifically, human neuroimaging with magnetic resonance (MRI). However, the uncertainty estimation associated these methods been largely limited to application of generic techniques (e.g., Monte Carlo dropout) that do not exploit peculiarities problem domain, particularly spatial modeling. Here, we propose a principled way propagate uncertainties...