A5041569484- Radiomics and Machine Learning in Medical Imaging
- AI in cancer detection
- Lung Cancer Diagnosis and Treatment
- Face Recognition and Perception
- Visual Attention and Saliency Detection
- Image Retrieval and Classification Techniques
- Surfactants and Colloidal Systems
- Carbon Nanotubes in Composites
- Graphene and Nanomaterials Applications
- Advanced Image and Video Retrieval Techniques
- Analytical Chemistry and Chromatography
- Electrochemical sensors and biosensors
- Monoclonal and Polyclonal Antibodies Research
- Advanced Nanomaterials in Catalysis
- Conducting polymers and applications
- Lipid Membrane Structure and Behavior
- Analytical chemistry methods development
- Graphene research and applications
- Photodynamic Therapy Research Studies
- Tea Polyphenols and Effects
- Wildlife-Road Interactions and Conservation
- Nanoplatforms for cancer theranostics
- Energy Load and Power Forecasting
- Pickering emulsions and particle stabilization
- Automated Road and Building Extraction
Nankai University
2020-2021
Columbia University Irving Medical Center
2013-2016
Anhui Agricultural University
2014
University of Oklahoma
2006-2010
Memorial Sloan Kettering Cancer Center
2009
Chinese Academy of Sciences
2009
Shanghai Institute of Technical Physics
2005-2007
University of Mississippi
2003-2006
Purpose: Lung lesions vary considerably in size, density, and shape, can attach to surrounding anatomic structures such as chest wall or mediastinum. Automatic segmentation of the poses a challenge. This work communicates new three‐dimensional algorithm for wide variety lesions, ranging from tumors found patients with advanced lung cancer small nodules detected screening programs. Methods: The authors’ uniquely combines image processing techniques marker‐controlled watershed, geometric...
Extracting road graphs from aerial images automatically is more efficient and costs less than field acquisition. This can be done by a post-processing step that vectorizes segmentation predicted CNN, but imperfect predictions will result in with low connectivity. On the other hand, iterative next move exploration could construct better connectivity, often focuses on local information does not provide precise alignment real road. To enhance connectivity while maintaining between graph road,...
CNN-based salient object detection (SOD) methods achieve impressive performance. However, the way semantic information is encoded in them and whether they are category-agnostic less explored. One major obstacle studying these questions fact that SOD models built on top of ImageNet pre-trained backbones which may cause leakage feature redundancy. To remedy this, here we first propose an extremely light-weight holistic model tied to task can be freed from classification trained scratch, then...
In this study, we describe the use of a sodium cholate suspension-dialysis method to adsorb redox enzyme glucose oxidase (GOX) onto single-walled carbon nanotubes (SWNT). By method, solutions dispersed and debundled SWNTs were prepared that remained stable for 30 days which retained 75% native enzymatic activity. We also demonstrate GOX-SWNT conjugates can be assembled into amperometric biosensors with poly[(vinylpyridine)Os(bipyridyl)2Cl2+/3+] polymer (PVP-Os) through layer-by-layer (LBL)...
The diversity of lung nodules poses difficulty for the current computer-aided diagnostic (CAD) schemes nodule detection on computed tomography (CT) scan images, especially in large-scale CT screening studies. We proposed a novel CAD scheme based hybrid method to address challenges diverse nodules.The this paper integrates several existing and widely used algorithms field detection, including morphological operation, dot-enhancement Hessian matrix, fuzzy connectedness segmentation, local...
Breast tumor volume measured on MRI has been used to assess response neoadjuvant chemotherapy. However, accurate and reproducible delineation of breast lesions can be challenging, since the may have complicated topological structures heterogeneous intensity distributions. In this article, authors present an advanced computerized method semiautomatically segment volumes T1‐weighted, contrast‐enhanced MRI. The starts with manual selection a region interest (ROI) that contains lesion segmented...
The objective of this study is to demonstrate the adsorption horseradish peroxidase (HRP) on single-walled carbon nanotubes (SWNTs) using sodium cholate suspension–dialysis method and determine effect HRP biological activity UV–vis–NIR spectra SNWTs. results indicate that in a stable SWNT–protein suspension with complete retention enzymatic adsorbed also substantial fraction NIR absorption at 980 nm. loading protein SWNTs high, overall yield preparing high. This process promising for...
Radiomics utilizes a large number of image-derived features for quantifying tumor characteristics that can in turn be correlated with response and prognosis. Unfortunately, extraction analysis such image-based is subject to measurement variability bias. The challenge radiomics particularly acute Positron Emission Tomography (PET) where limited resolution, high noise component related the stochastic nature raw data, wide variety reconstruction options confound quantitative feature metrics....
Chlorothalonil (CTL) is a broad-spectrum fungicide. Photodegradation main degradation pathway of CTL in water. Because the high aquatic toxicity and its metabolite 4-hydroxy (CTL-OH), it significant to develop an effective method degrade but without formation CTL-OH. Epigallocatechin gallate (EGCG) abundant tea byproduct has more than 100-fold reducing power vitamin C. The present study reports photosensitization effects EGCG on photodegradation water under sunlight artificial lights....
Traditionally, Picture Archiving and Communication Systems (PACS) use textual-based retrieval, which have their limitations. General-purposed content-based image retrieval (CBIR) systems often do not perform well in medical images are integrated with PACS Radiology Information System (RIS). In this presentation we design a CBIR system that is RIS, by using user-supplied query to retrieve similar from get corresponding reports RIS. We also employ ACR index for radiological diagnosis reduce...
A fractal analysis, which takes into account the effect of surface heterogeneity brought about by ligand immobilization on reaction kinetics in plasmon resonance (SPR) biosensors, is presented. The binding and dissociation estrogen receptors (ERs), ERa ER/spl alpha/ beta/, solution to different ligands immobilized SPR biosensor analyzed within framework. made quantitative using a single number, dimension D/sub f/. analysis provides physical insights these compounds, particularly endocrine...
Salient object detection models often demand a considerable amount of computation cost to make precise prediction for each pixel, making them hardly applicable on low-power devices. In this paper, we aim relieve the contradiction between and model performance by improving network efficiency higher degree. We propose flexible convolutional module, namely generalized OctConv (gOctConv), efficiently utilize both in-stage cross-stages multi-scale features, while reducing representation...
Carbon nanotubes have a great potential in the biomedical applications. To use carbon treatment of cancer, we synthesized an immunologically modified single-walled nanotube (SWNT) using novel immunomodifier, glycated chitosan (GC), as effective surfactant for SWNT. This new composition SWNT-GC was stable due to strong non-covalent binding between SWNT and GC. The structure is presented this report. photothermal effect investigated under irradiation near-infrared laser. retained optical...
Accurate day-ahead electricity price forecasting plays a vital role in power economic dispatch, system maintenance, and planning management, which can provide references for market participants dispatch control, energy trading, risk management. However, due to the influence of various uncertain factors, prices often exhibit nonlinearity, randomness, volatility, adding technical difficulties study forecasting. In order solve these difficulties, new hybrid deep learning method based on CNN...