A5101885804- Infrared Target Detection Methodologies
- Advanced Measurement and Detection Methods
- Remote-Sensing Image Classification
- Infrastructure Maintenance and Monitoring
- Advanced Neural Network Applications
- Occupational Health and Safety Research
- Infrared Thermography in Medicine
- Optical Systems and Laser Technology
- Computational Drug Discovery Methods
- Calibration and Measurement Techniques
- Advanced Image Fusion Techniques
- Medical Image Segmentation Techniques
- Advanced Semiconductor Detectors and Materials
- CCD and CMOS Imaging Sensors
- Advanced SAR Imaging Techniques
- Medical Imaging and Analysis
- Ocular and Laser Science Research
- Traffic and Road Safety
- Face and Expression Recognition
- Natural product bioactivities and synthesis
- Pharmacological Effects of Natural Compounds
- Video Surveillance and Tracking Methods
- Medical Imaging Techniques and Applications
- Imbalanced Data Classification Techniques
- Restless Legs Syndrome Research
Henan Provincial People's Hospital
2022-2025
Hunan University
2024-2025
Zhengzhou University
2011-2024
Jiangxi Science and Technology Normal University
2022-2024
Lanzhou University
2024
National Space Science Center
2017-2023
Chinese Academy of Sciences
2017-2023
Anhui Medical University
2022
Southern Medical University
2022
State Key Laboratory of Silkworm Genomic Biology
2022
A new deep learning method based on the target gray-level distribution constraint mechanism model is proposed to solve infrared dim small detection problem in complex environment. First, uneven of positive and negative samples, designed smoothness operator used suppress background enhancement by measuring difference their features 1D 2D gradient. Second, an network dense feature fusion, namely DFFIR-net network, proposed. The enhances expression targets integrating original Also, alleviates...
Molybdate-based inorganic–organic hybrid disks with a highly ordered layered structure were synthesized via an acid–base reaction of white molybdic acid (MoO3·H2O) n-octylamine (C8H17NH2) in ethanol at room temperature. The thermal treatment the as-obtained molybdate-based 550 °C air led to formation orthorhombic α-MoO3 nanoplates. X-ray diffraction (XRD), scanning electron microscopy (SEM), transmission (TEM), analysis (TG–DTA), Fourier-transform infrared (FT–IR) spectra, Raman and...
Tower cranes are the most commonly used large-scale equipment on construction site. Because workers can't always pay attention to environment at top of head, it is often difficult avoid accidents when heavy objects fall. Therefore, safety such as struck-by occurs. In order address crane issue, this research recorded video data by a tower camera, labeled pictures, and operated image recognition with MASK R-CNN method. Furthermore, The RGB color extraction was performed identified mask layer...
In the complex background, contrast and signal-to-noise (SNR) ratio of infrared (IR) small target are low. Therefore, traditional IR detection algorithms difficult to achieve good performance when characteristics targets sparse. To solve this problem, an network with generate label feature mapping (GLFM)-net is proposed in letter. First, GLFM-net model, a scale adaptive extraction for sparse features extraction, then, multilayer joint upsampling background suppression. Based on results dim...
Construction jobs are more labor intensive than other industrial jobs. Safety problems caused by overworked bodies common, and the supervision of construction workers is always flawed. In China, piecework has long been common way to evaluate workers' workloads, because it inconvenient obtain direct indicators. To improve this situation, paper proposes a method based on smartphone sensor acquisition concept intensity workloads. A application platform was created effectively measure so that...
The construction industry is facing unique problems in accident prevention. existing management method for detecting workers’ unsafe behaviors and states of objects relies primarily on manual monitoring, which does not only consume large amounts time money but also cannot cover all workers the entire site. Meanwhile, perception being at risk injury decreases when they are concentrated a crowded noisy environment. In this case, it difficult them to take essential measures protect themselves...
Aiming at the problems of large parameters and low detection speed existing remote sensing road network model, a lightweight (LRSR-net) model is proposed in this paper. In feature encoding stage LRSR-net, an efficient target extraction module dilated joint convolution proposed. This can eliminate loss caused by pooling layer, so that use small amount to learn characteristics targets. addition, decoding stage, damage data structure upsampling operation, The effectively map targets suppress...
Dense pixel-wise labeling of large-scale remote sensing images (RSI) is very time-consuming, while sparse labels (i.e., points, scribbles, or blocks) can be an efficient way to reduce costs. Most existing label-based methods adopt only one type label for image segmentation, which cannot reflect the complex land covers in RSI training model, thus leading inferior segmentation performance. We observe that with different shapes and complexity optimally represented by labels. Inspired this...
Intractable sparse feature extraction, over-weight model size, and limited training samples are currently bewildering in infrared dim small target detection, which not adequately addressed by current state-of-the-art (SOTA) methods. Here, to synchronously address these issues, a dense multi-level extraction fusion network (DMEF-net) is designed, mainly consisting of two modules: context Gaussian saliency module (TCGS) structure (MLDF). Inspired the physical thermal diffusion human visual...
The success of deep learning methods heavily relies on the availability adequate samples. However, in task infrared small target detection (ISTD), lack high-quality training samples is a challenging problem due to confidentiality application field and difficulty labeling. This limitation often leads suboptimal performance convolutional neural networks (CNNs). To address this challenge, we propose an adaptive labeling strategy ISTD network called multi-criteria decision (MCDNet) achieve...
The detection of dim and small space-targets is crucial in space situational awareness missions; however, low signal-to-noise ratio (SNR) targets complex backgrounds pose significant challenges to such detection. This paper proposes a space-target framework comprising network k-means clustering target centroid positioning method. performs three-dimensional convolution an input star image sequence learn the motion features target, reduces interference noise using soft thresholding module,...
This letter presents an approach for the detection of moving point targets on high-frame-rate image sequences with low spatial resolution and SNR based higher order statistical theory. We propose a novel method analyzing time-domain evolution data distinguishing between background target in situations when signal is swamped by noise. Our formulated to detect transient unknown scale arrival time noisy background. proposed bispectrum-based model characterize temporal behavior pixels, ability...
Oxidative stress is the redox imbalance state of organisms that involves in a variety biological processes diseases. Limonium aureum (L.) Hill. an excellent wild plant resource northern China, which has potential application value for treating oxidative stress. However, there are few studies focused on antioxidant effect and related mechanism L. . Thus, present study combining systematic network pharmacology molecular biology aimed to investigate effects explore its underlying anti-oxidation...
Forest fires pose a catastrophic threat to Earth’s ecology as well threaten human beings. Timely and accurate monitoring of forest can significantly reduce potential casualties property damage. Thus, address the aforementioned problems, this paper proposed an unmanned aerial vehicle (UAV) based on lightweight fire recognition model, Fire-Net, which has multi-stage structure incorporates cross-channel attention following fifth stage. This is enable model’s ability perceive features at various...
With the rapid development of Internet, especially extensive application deep learning, change detection has been successfully applied in many fields, but pursuit greater feature information content led to an increase memory and computing power requirements. To solve this problem, a Lightweight Siamese attention Residual Network (LSRNet) is proposed paper reduce computational requirements, embed Fast Small Attention (SFSAttention) filter out with less relevance, then use fusion channel...
Infrared (IR) dim small target detection in heavy noise has always been a challenging research task. Detection methods based on energy accumulation have achieved effective results. However, the general inaccuracy of distribution model for subpixel motion weakens performance these methods. The discretization process is described detail, and effect two-dimensional (2D) spatial introduced. Based analysis motion, novel 2D (NESDSM) proposed. NESDSM well describes variation moving target....
To solve the problems of slow detection speed and poor robustness existing infrared (IR) small target methods in complex environments, a lightweight model MiniIR-net is proposed this letter. In model, to reduce number parameters required for fitting, multiscale context feature extraction (TCVE) module enrich expression target. addition, improve mapping capability MiniIR-net, upsampling network by fusing deep shallow features designed. process upsampling, uses with different depths make up...
Gustatory systems in phytophagous insects are used to perceive feeding stimulants and deterrents, involved insect decisions feed on particular plants. During the process, gustatory receptors (Grs) can recognize diverse phytochemicals provide a molecular basis for taste perception. The silkworm, as representative Lepidoptera species, has developed strong preference mulberry leaves. mulberry-derived flavonoid glycoside, isoquercetin, is required induce behaviours. However, corresponding Grs...
Infrared (IR) small target detection in a complex environment has become hot topic. Due to the dim intensity and scale of strong background clutter, existing algorithms cannot achieve satisfactory results, namely poor performance real-time performance. In this letter, new algorithm based on smoothness measure thermal diffusion (STD) is proposed. First, proposed extract features targets suppress clutter background. Then, we improve equation design operator target’s feature. The will further...