A5088998219- Advanced X-ray and CT Imaging
- Medical Imaging Techniques and Applications
- Medical Imaging and Analysis
- Hip and Femur Fractures
- Adversarial Robustness in Machine Learning
- Artificial Intelligence in Healthcare and Education
- Advanced Neural Network Applications
- Domain Adaptation and Few-Shot Learning
- COVID-19 diagnosis using AI
- Genetics, Aging, and Longevity in Model Organisms
- Radiomics and Machine Learning in Medical Imaging
- Epigenetics and DNA Methylation
- EEG and Brain-Computer Interfaces
- Neurological disorders and treatments
- Advanced MRI Techniques and Applications
- Genetics and Neurodevelopmental Disorders
- Spaceflight effects on biology
- Circadian rhythm and melatonin
- Advanced Computing and Algorithms
- Advanced Algorithms and Applications
- RNA modifications and cancer
- Neuroethics, Human Enhancement, Biomedical Innovations
- Integrated Circuits and Semiconductor Failure Analysis
- Advanced Sensor and Control Systems
University of California, Los Angeles
2022-2024
Shanghai Maritime University
2023
University of California, Berkeley
2019-2021
University of California System
2021
University of California, San Francisco
2021
Using DNA methylation profiles (
Purpose To investigate the feasibility of automatic identification and classification hip fractures using deep learning, which may improve outcomes by reducing diagnostic errors decreasing time to operation. Materials Methods Hip pelvic radiographs from 1118 studies were reviewed, 3026 hips labeled via bounding boxes classified as normal, displaced femoral neck fracture, nondisplaced intertrochanteric previous open reduction internal fixation, or arthroplasty. A learning–based object...
Abstract Claims surrounding exceptional longevity are sometimes disputed or dismissed for lack of credible evidence. Here, we present three DNA methylation-based age estimators (epigenetic clocks) verifying claims centenarians. The centenarian clocks were developed based on n = 7039 blood and saliva samples from individuals older than 40, including 184 centenarians, 122 semi-supercentenarians (aged 105 +), 25 supercentenarians 110 +). oldest individual was 115 years old. Our most accurate...
To gain insight into how researchers of aging perceive the process they study, we conducted a survey among experts in field. While highlighting some common features aging, exposed broad disagreement on foundational issues. What is aging? causes it? When does it begin? constitutes rejuvenation? Not only was there no consensus these and other core questions, but none questions received majority opinion-even regarding need for itself. Despite many believing understand their understanding...
Deep Learning (DL) has shown potential in accelerating Magnetic Resonance Image acquisition and reconstruction. Nevertheless, there is a dearth of tailored methods to guarantee that the reconstruction small features achieved with high fidelity. In this work, we employ adversarial attacks generate synthetic perturbations, which are difficult reconstruct for trained DL network. Then, use robust training increase network’s sensitivity these encourage their Next, investigate generalization said...
Emotion recognition has received great attention in the fields of emotional computing and human-computer interaction. Due to limited number EEG signal samples individual differences, efficient feature extraction fusion brain dynamic information have become key issues. Inspired by Google Network, a new convolutional neural network called TFSNet (Time Frequency Space Network) been proposed. The includes temporal layer, frequency domain spatial block optimization layer. Unlike traditional...
The use of semantic segmentation for masking and cropping input images has proven to be a significant aid in medical imaging classification tasks by decreasing the noise variance training dataset. However, implementing this approach with classical methods is challenging: cost obtaining dense high, precise area that most crucial task difficult determine a-priori. We propose novel joint-training deep reinforcement learning framework image augmentation. A network, weakly supervised policy...
Purpose: Hip fractures are a common cause of morbidity and mortality. Automatic identification classification hip using deep learning may improve outcomes by reducing diagnostic errors decreasing time to operation. Methods: pelvic radiographs from 1118 studies were reviewed 3034 hips labeled via bounding boxes classified as normal, displaced femoral neck fracture, nondisplaced intertrochanteric previous ORIF, or arthroplasty. A learning-based object detection model was trained automate the...
Deep Learning (DL) has shown potential in accelerating Magnetic Resonance Image acquisition and reconstruction. Nevertheless, there is a dearth of tailored methods to guarantee that the reconstruction small features achieved with high fidelity. In this work, we employ adversarial attacks generate synthetic perturbations, which are difficult reconstruct for trained DL network. Then, use robust training increase network's sensitivity these encourage their Next, investigate generalization said...
Temporarily implanted electrocorticographic (ECoG) electrodes provide cortical electrical activity data during deep brain stimulation (DBS) surgery. Accurate localization of ECoG in 3D space is required for the recorded to be useful research, but these are only intraoperatively. During surgery, imaging modality available fluoroscopy, which produces a 2D rather than image. Current approaches require postoperative (postop) computed tomography (CT) volumes register with intraoperative...