- Phonetics and Phonology Research
- Speech Recognition and Synthesis
- Speech and Audio Processing
- Image Enhancement Techniques
- Advanced Image Processing Techniques
- Advanced Vision and Imaging
- Image Processing Techniques and Applications
- Microwave Dielectric Ceramics Synthesis
- Ferroelectric and Piezoelectric Materials
- Optical measurement and interference techniques
- Image and Signal Denoising Methods
- Heat Transfer and Optimization
- Industrial Vision Systems and Defect Detection
- Advanced Image Fusion Techniques
- Voice and Speech Disorders
- Textile materials and evaluations
- Environmental Impact and Sustainability
- Heat Transfer and Boiling Studies
- Welding Techniques and Residual Stresses
- Water Quality Monitoring and Analysis
- Multiferroics and related materials
- Image and Object Detection Techniques
- Turbomachinery Performance and Optimization
- Copper-based nanomaterials and applications
- Advanced X-ray and CT Imaging
Jiangsu University of Science and Technology
2025
Hunan University of Science and Technology
2024
China Jiliang University
2024
Northwest Normal University
2024
Central South University
2020-2023
Huazhong University of Science and Technology
2023
Hefei University of Technology
2021-2022
Science and Technology Department of Sichuan Province
2022
Chengdu University of Information Technology
2022
State Grid Corporation of China (China)
2022
X-ray image plays an important role in manufacturing industry for quality assurance, because it can reflect the internal condition of weld region. However, shape and scale different defect types vary greatly, which makes challenging model to detect defects. In this article, we propose a detection method based on convolution neural network (CNN), namely, lighter faster YOLO (LF-YOLO). particular, reinforced multiscale feature (RMF) module is designed implement both parameter-based...
Fabric quality plays a crucial role in modern textile industry processes. How to detect fabric defects quickly and effectively has become the main research goal of researchers. The You Only Look Once (YOLO) series networks have maintained dominant position field target detection. However, detecting small-scale objects, such as tiny targets defects, is still very challenging task for YOLOv4 network. To address this challenge, paper proposed an improved detection algorithm: using combined data...
Edge preserving filter is the basis of many computational photography and image processing. This can be achieved by global optimization method or local filtering method. Generally, results methods are better than that methods, usually run much faster methods. In this paper, a globally optimized called iterative self-guided (isGIF) extended based on assumptions guided (GIF), which produce high-quality edge-preserving using input itself as guidance image. Some comparisons with other edge-aware...
ABSTRACT Mg‐doped Cu 2 ZnSnS 4 films were prepared by sol–gel method with varying Mg concentration. The effects of concentration on the phase structure, optical properties, and electrical properties investigated. incorporation did not alter structure but caused a shift diffraction peaks to larger angles slight reduction in bandgap as increased. Sample S3‐4% exhibited higher lattice strain, whereas sample S4‐6% presented lowest strain value 0.63. XPS results demonstrate that chemical valence...
Biomolecules exhibit conformational fluctuations near equilibrium states, inducing uncertainty in various biological properties a dynamic way. We have developed general method to quantify the of target induced by fluctuations. Using generalized polynomial chaos (gPC) expansion, we construct surrogate model property with respect varying states. also propose increase sparsity gPC expansion defining set "active space" random variables. With increased sparsity, employ compressive sensing...
ZnO-doped Mg2TiO4–MgTiO3–CaTiO3 microwave dielectric ceramics were successfully prepared by the reaction sintering route. The compact samples consisted of MgTiO3, Mg2TiO4 and CaTiO3, which was confirmed X-ray diffraction energy-dispersive spectra. ZnO efficiently lowered temperature promoted densification, as well improvements in constant quality factor. At level = 1 wt-%, exhibited optimum properties: a 20.3, high factor 64,740 GHz (at 9.9 GHz) near-zero coefficient resonant frequency (–1.3...
The present work numerically studies the thermal characteristics of a staggered double-layer microchannel heat sink (DLMCHS) with an offset between upper layer microchannels and lower in width direction, investigates effects inlet velocity geometric parameters including two layers microchannels, vertical rib thickness aspect ratio on resistance DLMCHS. found that DLMCHS increases increasing value when is small, but decreases firstly then as large enough. Furthermore, Thus, for small...
The purpose of texture smoothing is to preserve the prominent structure in image while salient texture. However, existing methods are difficult achieve a satisfactory balance between filtering out textures and preserving weak edge structures small structures. To this end, we propose structure-preserving method via scale-aware bilateral total variation. First, joint filter introduced construct window variation, combined with regularizer called variation formed, which accurately quantifies...
Abstract BaTi 4 O 9 microwave dielectric ceramics were prepared by reaction sintering method using Ba CO 3 and TiO 2 as raw materials. The phase evolution the chemical reactions proposed based on X‐ray diffraction results with temperature. microstructure characteristics observed scanning electron microscopy energy dispersive spectrometer. compact a single of could be successfully method, exhibiting optimum properties: constant 36.9, high quality factor 52 735 (at 7.5 GH z), near zero...
We model the amplitude envelope of a speech signal as kinematic system and calculate its basic parameters: displacement, velocity, acceleration. Such captures smoothed fluctuation pattern over time, illustrating how energy is distributed across signal. Although pulmonic air pressure primary source speech, modulation largely determined by articulatory behaviors, especially mandible lip movements. Therefore, there should be correspondence between kinematics articulator kinematics. Previous...
Intensity contours of speech signals were sub-divided into positive and negative dynamics. Positive dynamics defined as the speed increases in intensity from amplitude troughs to subsequent peaks, decreases peaks troughs. Mean, standard deviation, sequential variability measured for both each sentence. Analyses showed that measures separately classified between-speaker was largely explained by This suggests parts signal where syllable are more speaker-specific. Idiosyncratic articulation may...
In this paper, we propose a novel algorithm for analysis-based sparsity reconstruction. It can solve the generalized problem by structured regularization with an orthogonal basis and total variation regularization. The proposed is based on iterative reweighted least squares (IRLS) model, which further accelerated preconditioned conjugate gradient method. convergence rate of almost same as that traditional IRLS algorithms, is, exponentially fast. Moreover, specifically devised preconditioner,...
The opening-closing alternations of the mouth were viewed as articulatory basis speech rhythm. Such cycles have been observed to highly correlate with intensity curve signal. Analysis variability in English monolingual children and adults revealed that (1) showed significantly smaller than children, (2) decreased from intermediate-aged older children. Maturation motor control is likely be main reason for reduced cycles, hence
China has a considerable impact on climate change, as it is the world’s largest developing country. Its carbon emissions are also global concern. This research employed multi-region input–output model to estimate local emission effect and regional spillover caused by demand for intermediate final products, analyze interregional flow of emissions, identify main industries causing domestic in from 2007 2012. A structural path analysis was used evaluate contribution different factors types...
Accurately measuring the velocity of molten iron at taphole a blast furnace is great significance for safe tapping and reflecting operating state furnace. The presence dust-induced brightness changes weakly textured areas on surface seriously affects accuracy field reconstructed by traditional methods. This paper presents local multi-feature correction using multi-stage filtered high-speed camera to achieve high-precision measurement iron. First, two sets features were extracted according...
In this study, three intensity metrics (ΔS.dB, VarcoS.dB and nPVI.dB) devised on the basis of well-known durational were tested both L1 (English Mandarin) L2 (L2 English). The results suggested that they effective in distinguishing “stress-timed” English from perceptually “syllable-timed” Mandarin (by speakers). These break impasse although varieties similar measurements, different rhythmicity. Therefore, it is advised be applied future rhythm research.
One of the most troublesome problems in development a component-based engine model is compressor modeling because strong dependence its performance on rotational speed and treatment characteristic curves plays significant role simulation gas turbine for analysis off-design thus it crucial to describe map exactly. Usually part including on-design operating point are known during actual simulation, reasonable interpolation extrapolation have be done order make good use flow characteristics...