A5003767673- Retinal Imaging and Analysis
- Retinal Diseases and Treatments
- Advanced Neural Network Applications
- Medical Image Segmentation Techniques
- Glaucoma and retinal disorders
- Brain Tumor Detection and Classification
- Image Enhancement Techniques
- Machine Learning and Data Classification
- Acute Ischemic Stroke Management
- Energy Load and Power Forecasting
- Medical Imaging Techniques and Applications
- Neural Networks and Applications
- COVID-19 diagnosis using AI
- Retinal and Optic Conditions
- Evolutionary Algorithms and Applications
- Global trade, sustainability, and social impact
- Stock Market Forecasting Methods
- Radiomics and Machine Learning in Medical Imaging
- Optical Coherence Tomography Applications
- Chinese history and philosophy
- Reservoir Engineering and Simulation Methods
- Metaheuristic Optimization Algorithms Research
- Anomaly Detection Techniques and Applications
- Image and Signal Denoising Methods
- Forecasting Techniques and Applications
University of Toronto
2023
Friedrich-Alexander-Universität Erlangen-Nürnberg
2018-2022
Max Planck Institute for the Science of Light
2020-2021
Renmin University of China
2018
Interpretability is crucial for machine learning in many scenarios such as quantitative finance, banking, healthcare, etc. Symbolic regression (SR) a classic interpretable method by bridging X and Y using mathematical expressions composed of some basic functions. However, the search space all possible grows exponentially with length expression, making it infeasible enumeration. Genetic programming (GP) has been traditionally commonly used SR to optimal solution, but suffers from several...
In this paper, we reformulate the conventional 2-D Frangi vesselness measure into a pre-weighted neural network ("Frangi-Net"), and illustrate that Frangi-Net is equivalent to original filter. Furthermore, show that, as network, trainable. We evaluate proposed method on set of 45 high resolution fundus images. After fine-tuning, observe both qualitative quantitative improvements in segmentation quality compared measure, with an increase up $17\%$ F1 score.
: Every financial crisis has caused a dual shock to the global economy. The shortage of market liquidity, such as default in debt and bonds, led spread bankruptcies, Lehman Brothers 2008. Using data for ETFs S&P 500, Nasdaq 100, Dow Jones Industrial Average collected from Yahoo Finance, this study implemented Deep Learning, Neuro Network, Time-series analyze trend American Stock Market post-COVID-19 period. LSTM model Network predict future trend, which suggests US stock keeps falling...
Abstract Purpose With the recent development of deep learning technologies, various neural networks have been proposed for fundus retinal vessel segmentation. Among them, U-Net is regarded as one most successful architectures. In this work, we start with simplification U-Net, and explore performance few-parameter on task. Methods We firstly modify model popular functional blocks additional resolution levels, then switch to exploring limits compression network architecture. Experiments are...
Retinal vessel segmentation is an essential step for fundus image analysis. With the recent advances of deep learning technologies, many convolutional neural networks have been applied in this field, including successful U-Net. In work, we firstly modify U-Net with functional blocks aiming to pursue higher performance. The absence expected performance boost then lead us dig into opposite direction shrinking and exploring extreme conditions such that its maintained. Experiment series simplify...
China entered the Nagoya Protocol on September 6, 2016, signaling its establishment of regime regarding protection genetic resources and associated traditional knowledge as well Access Benefit Sharing system (ABS) at both national international levels.It is stated that purpose defensive to prevent bio-piracy while constructive facilitate benefit sharing.Based years field research, we reveal in this study a user De'ang acid tea, which with local knowledge, has applied for patent without...
Fundus photography and Optical Coherence Tomography Angiography (OCT-A) are two commonly used modalities in ophthalmic imaging. With the development of deep learning algorithms, fundus image processing, especially retinal vessel segmentation, has been extensively studied. Built upon known operator theory, interpretable network pipelines with well-defined modules have constructed on images. In this work, we firstly train a modularized pipeline for task segmentation database DRIVE. The...
Deep neural networks have achieved tremendous success in various fields including medical image segmentation. However, they long been criticized for being a black-box, that interpretation, understanding and correcting architectures is difficult as there no general theory deep network design. Previously, precision learning was proposed to fuse traditional approaches. constructed this way benefit from the original known operator, fewer parameters, improved interpretability. do not yield...
Comparison of microvascular circulation on fundoscopic images is a non-invasive clinical indication for the diagnosis and monitoring diseases, such as diabetes hypertensions. The differences between intra-patient can be assessed quantitatively by registering serial acquisitions. Due to variability (i.e. contrast, luminosity) anatomical changes retina, registration fundus remains challenging task. Recently, several deep learning approaches have been proposed register in an end-to-end fashion,...