A5101911538- Target Tracking and Data Fusion in Sensor Networks
- Inertial Sensor and Navigation
- Fault Detection and Control Systems
- Astronomical Observations and Instrumentation
- Image Enhancement Techniques
- 2D Materials and Applications
- Quantum Information and Cryptography
- Astronomy and Astrophysical Research
- Adversarial Robustness in Machine Learning
- GNSS positioning and interference
- MXene and MAX Phase Materials
- Mechanical and Optical Resonators
- Perovskite Materials and Applications
- Indoor and Outdoor Localization Technologies
- Advanced Measurement and Detection Methods
- Machine Learning in Materials Science
- Inorganic Chemistry and Materials
- Color perception and design
- Video Surveillance and Tracking Methods
- TiO2 Photocatalysis and Solar Cells
- Advanced Image Processing Techniques
- Cyclone Separators and Fluid Dynamics
- Bone Tumor Diagnosis and Treatments
- Advanced Neural Network Applications
- Emotion and Mood Recognition
Tianjin Renai College
2021-2025
Air Force Engineering University
2015-2025
Chaozhou Central Hospital
2024-2025
Jilin University
2024-2025
Qiqihar University
2006-2024
Chongqing University
2024
Shantou University
2014-2023
Harbin University of Science and Technology
2023
Shantou Central Hospital
2014-2023
Zhongshan People's Hospital
2022-2023
This work reviews the results of NTIRE 2021 Challenge on Non-Homogeneous Dehazing. The proposed techniques and their have been evaluated a novel dataset that extends NH-Haze datset. It consists additional 35 pairs real haze free nonhomogeneous hazy images recorded outdoor. has introduced in outdoor scenes by using professional setup imitates conditions scenes. 327 participants registered challenge 23 teams competed final testing phase. solutions gauge state-of-the-art image dehazing.
This paper reviews the second NTIRE challenge on image dehazing (restoration of rich details in hazy image) with focus proposed solutions and results. The training data consists from 55 images (with dense haze generated an indoor or outdoor environment) their corresponding ground truth (haze-free) same scene. has been produced using a professional haze/fog generator that imitates real conditions scenes. evaluation comparison dehazed images. process was learnable through provided pairs...
Virtual high-throughput screening (VHTS) with density functional theory (DFT) and machine-learning (ML)-acceleration is essential in rapid materials discovery. By necessity, efficient DFT-based workflows are carried out a single approximation (DFA). Nevertheless, properties evaluated different DFAs can be expected to disagree for cases challenging electronic structure (e.g., open-shell transition-metal complexes, TMCs) which most needed accurate benchmarks often unavailable. To quantify the...
Background Patients with end-stage renal disease (ESRD) have increased insulin resistance (IR). The triglyceride glucose (TyG) index is a marker of IR and associated the prognosis patients on maintenance hemodialysis (MHD). aim this study was to examine relationship between TyG risk death in MHD. Methods In retrospective cohort analysis 368 MHD over an 11-year period from July 1, 2012, June 30, 2023, other baseline characteristics were measured at beginning treatment. A regression model used...
Abstract The burgeoning demand for efficient photoelectric devices has ignited a fervent exploration of strategies regulating their performance. Herein, the enhancement and polarity switching in photoresponse MoS 2 photovoltaic with asymmetric electrodes are showcased by employing pressure band engineering strategy. Specifically, Pt/MoS /Ag setup escalates from A W⁻¹ at 0.5 GPa to 4 2.3 GPa, showcasing twofold enhancement, while displaying successive decrease under higher pressure....
Abstract The effective modification of photovoltaic activities in photoelectric devices is significant for enhancing energy conversion efficiency. Here, the pressure‐tunable properties SbI 3 device with Au/Pt asymmetric electrodes are demonstrated. At initial pressure, exhibits an anticipated self‐driven photoresponse characteristic at zero external bias, and a enhancement photocurrent response observed increasing bias to −0.2 V, nearly two orders magnitude higher than value. With tunable...
Aiming at the problem of insufficient feature extraction and low recognition rate traditional convolutional neural network in facial expression recognition, a multi-layer algorithm based on three-channel (HFT-CNN) was proposed. In network, one channel trained local image eyes eyebrows extracted from image, another mouth last tained whole image. order to better extract features three channel, higher-order tensor singular value decomposition used decompose convolution kernel remove...
Since haze degrades an image including contrast decreasing and color lost, which has a negative effect on the subsequent object detection recognition. single dehazing is challenging visual task. Most existing methods are not robust to uneven haze. In this paper, we developed adaptive distillation network solve dehaze problem with non-uniform haze, does rely physical scattering model. The proposed model consists of two parts: module multi-scale enhancing module. block reassigns channel...
Aiming at the nonlinearity, nonstationarity and multi-component coupling characteristics of reciprocating compressor vibration signals, an integrated feature extraction method based on variational mode decomposition (VMD) multi-fractal detrended fluctuation analysis (MFDFA) is proposed for a fault diagnosis valve. Firstly, to eliminate noise interference, novel VMD with superior anti-interference performance was utilized obtain several components quasi-orthogonal band-limited intrinsic...
Deep learning techniques such as convolutional neural networks (CNNs) have been used in a wide range of fields due to their superior performance, e.g., image classification, autonomous driving and natural language processing. However, recent progress shows that deep models are vulnerable adversarial samples, which crafted by adding small perturbations on normal samples imperceptible human beings but can mislead the output incorrect results. Many attack proposed many detection methods...
<b><i>Introduction:</i></b> Primary brainstem hemorrhage (PBSH) is the most fatal subtype of intracerebral and associated with poor prognosis. We aimed to develop a prediction model for predicting 30-day mortality functional outcome in patients PBSH. <b><i>Methods:</i></b> reviewed records 642 consecutive first-time PBSH from three hospitals between 2016 2021. Multivariate logistic regression was used establish nomogram training cohort. Cutoff...
The recognition rate of person-independent facial expression is generally not high, which limits the practical application recognition. Aiming at this problem, paper analyzes reasons for low expression, and proposes a algorithm based on improved LBP (Local Binary Pattern) HOSVD (Higher-Order Singular Value Decomposition). has following contributions framework. In stage feature extraction, transient features extracted by LDP(Local Directional persistent CBP(Centralized are integrated to...