A5100395140- Retinal Imaging and Analysis
- Advanced Image and Video Retrieval Techniques
- Image Retrieval and Classification Techniques
- Retinal Diseases and Treatments
- Natural Language Processing Techniques
- Radiomics and Machine Learning in Medical Imaging
- Multimodal Machine Learning Applications
- Face and Expression Recognition
- Retinal and Optic Conditions
- Parallel Computing and Optimization Techniques
- Topic Modeling
- Video Surveillance and Tracking Methods
- Advanced Neural Network Applications
- AI in cancer detection
- Gait Recognition and Analysis
- Robotics and Sensor-Based Localization
- Medical Image Segmentation Techniques
- Cloud Computing and Resource Management
- Advanced X-ray and CT Imaging
- Advanced Image Processing Techniques
- Brain Tumor Detection and Classification
- Ophthalmology and Visual Impairment Studies
- Advanced Image Fusion Techniques
- Human Pose and Action Recognition
- Retinal and Macular Surgery
Dalian Institute of Chemical Physics
2025
Chinese Academy of Sciences
2025
Anhui University of Technology
2006-2024
Guangzhou University
2024
Nantong University
2024
Doheny Eye Institute
2021-2024
Sinosteel (China)
2024
Taizhou Fourth People's Hospital
2024
Taizhou People's Hospital
2024
China University of Petroleum, Beijing
2024
In video surveillance, the faces of interest are often small size. Image resolution is an important factor affecting face recognition by human and computer. this paper, we propose a new hallucination method using eigentransformation. Different from most proposed methods based on probabilistic models, views as transformation between different image styles. We use Principal Component Analysis (PCA) to fit input linear combination low-resolution images in training set. The high-resolution...
Automatic retrieval of face images from police mug-shot databases is critically important for law enforcement agencies. It can effectively help investigators to locate or narrow down potential suspects. However, in many cases, a photo image suspect not available and the best substitute often sketch drawing based on recollection an eyewitness. We present novel system using sketches. By transforming into sketch, we reduce difference between significantly, thus allowing effective matching two....
To meet the progressive requirements of advanced engineering materials with superior physicochemical performances, self-healing and injectable hydrogels (AD hydrogels) based on agarose pH-response were prepared through dynamic covalent Schiff-base linkages by simply mixing nontoxic agarose–ethylenediamine conjugate (AG-NH2) dialdehyde-functionalized polyethylene glycol (DF-PEG) solutions. The capabilities without any external stimulus are ascribed to between aldehyde groups DF-PEG amine...
Most CNN models exhibit two major flaws in hyper-spectral image (HSI) restoration tasks. First, limited high-dimensional HSI training examples exacerbate the difficulty of deep learning methods effective spatial and spectral representations. Second, existing CNN-based model local relations present limitations capturing long-range dependencies. In this paper, we customize a novel dual-stream Transformer (DSTrans) for restoration, which mainly consists attention feed-forward network....
Extended warranties (EWs) are significant source of revenue for capital-intensive products like automobiles. Such consist multiple subsystems, providing flexibility in EW customization, example, bundling a tailored set subsystems an contract. This, turn, enables the creation service menu with different contract options. From perspective third-party provider servicing fleet automobile brands, we develop novel model to jointly optimize design and pricing EWs order maximize profit....
Background/Objectives: Gastric cancer (GC) is a prevalent malignant tumor worldwide, with its pathological mechanisms largely unknown. Understanding the metabolic reprogramming associated GC crucial for prevention and treatment of this disease. This study aims to identify significant alterations in metabolites pathways related development GC. Methods: A liquid chromatography–mass spectrometry-based non-targeted metabolomics data acquisition was performed on paired tissues from 80 patients....
Subretinal drusenoid deposits (SDD), a recently recognized lesion associated with progression of age-related macular degeneration, were imaged adaptive optics scanning laser ophthalmoscopy (AO-SLO) and optical coherence tomography (AO-OCT).AO-SLO revealed distinct en face structure stage 3 SDD, showing hyporeflective annulus surrounded reflective core packed hyperreflective dots bearing superficial similarity to the photoreceptors in unaffected retina.However, AO-OCT suggested that speckled...
Preoperative evaluation of microvascular invasion (MVI) in small solitary hepatocellular carcinoma (HCC; maximum lesion diameter ≤ 3 cm) is important for treatment decisions.To apply gadolinium ethoxybenzyl diethylenetriamine pentaacetic acid (Gd-EOB-DTPA)-enhanced MRI to develop and validate a nomogram preoperative MVI HCC compare the effectiveness radiomics models based on different volumes interest (VOIs).Retrospective.A total 196 patients include 62 MVI-positive 134 MVI-negative were...
Deformable image registration plays a critical role in various tasks of medical analysis. A successful algorithm, either derived from conventional energy optimization or deep networks, requires tremendous efforts computer experts to well design carefully tune network architectures with respect data available for given task/scenario. This paper proposes an automated learning algorithm (AutoReg) that cooperatively optimizes both and their corresponding training objectives, enabling...
Deep learning (DL) shows its prosperity in a wide variety of fields. The development DL model is time-consuming and resource-intensive procedure. Hence, dedicated GPU accelerators have been collectively constructed into datacenter. An efficient scheduler design for such datacenter crucially important to reduce the operational cost improve resource utilization. However, traditional approaches designed big data or high performance computing workloads can not support fully utilize resources....
In this paper we examine the effectiveness of neural network sequence-to-sequence transduction in task transliteration generation.In year's shared evaluation submitted two systems into all tasks.The primary system was based on used for NEWS 2012 workshop, but augmented with an additional feature which generation probability from a network.The secondary model its own together simple beam search algorithm.Our results show that adding score as phrase-based statistical machine able to increase...
Recently, multimodal hashing techniques have received considerable attention due to their low storage cost and fast query speed for data retrieval. Many methods been proposed; however, there are still some problems that need be further considered. For example, of these just use a similarity matrix learning hash functions which will discard useful information contained in original data; them relax binary constraints or separate the process codes into two independent stages bypass obstacle...
Recent large-scale hierarchical classification tasks typically have tens of thousands classes on which the most widely used approach to multiclass classification--one-versus-rest--becomes intractable due computational complexity. The top-down methods are usually adopted instead, but they less accurate because so-called error-propagation problem in their classifying phase. To address this problem, paper proposes a meta-top-down method that employs metaclassification enhance normal procedure....
We present a technique to measure the rapid blood velocity in large retinal vessels with high spatiotemporal resolution. Red cell motion traces were non-invasively imaged using an adaptive optics near-confocal scanning ophthalmoscope at frame rate of 200 fps. developed software automatically. demonstrated ability profiles pulsatile flow maximum 95–156 mm/s arterioles diameter >100 µm. High-speed and high-resolution imaging increased dynamic range, enhanced sensitivity, improved accuracy...
Convolutional neural networks (CNNs) have dominated the research of hyperspectral image (HSI) classification, attributing to superior feature representation capacity. Patch-free global learning (FPGA) as a fast framework for HSI classification has received wide interest. Despite their promising results from perspective inference, recent works difficulty modeling spectral-spatial relationships with imbalanced samples. In this paper, we revisit encoder–decoder-based fully convolutional network...
Unsupervised word segmentation (UWS) can provide domain-adaptive for statistical machine translation (SMT) without annotated data, and bilingual UWS even optimize alignment. Monolingual approaches of explicitly modeling the probabilities words through Dirichlet process (DP) models or Pitman-Yor (PYP) have achieved high accuracy, but their counterparts only been carried out on small corpora such as basic travel expression corpus (BTEC) due to computational complexity. This paper proposes an...
The unmanned aerial vehicles (UAVs) have been widely used in various application fields, yet unauthorized use of UAVs raises great threats for restricted areas and public security. Therefore, it is urgently necessary to develop a practical anti-UAV target tracking technique. In this paper, we propose real-time anti-distractor infrared UAV tracker tasks, which employs global perception mechanism find candidate targets, then utilizes spatial-temporal information obtain the real target....