- Speech and Audio Processing
- Image and Signal Denoising Methods
- Underwater Acoustics Research
- Advanced Image Processing Techniques
- Advanced Computational Techniques and Applications
- Acoustic Wave Phenomena Research
- Sparse and Compressive Sensing Techniques
- Ultrasonics and Acoustic Wave Propagation
- Music and Audio Processing
- Structural Health Monitoring Techniques
- Image Processing Techniques and Applications
- Simulation and Modeling Applications
- Advanced Neural Network Applications
- Aerodynamics and Acoustics in Jet Flows
- Industrial Technology and Control Systems
- Advanced Adaptive Filtering Techniques
- Advanced Sensor and Control Systems
- Energy Efficient Wireless Sensor Networks
- Model Reduction and Neural Networks
- Advanced Wireless Communication Techniques
- Higher Education and Teaching Methods
- Water Systems and Optimization
- Speech Recognition and Synthesis
- Advanced Decision-Making Techniques
- Advanced Algorithms and Applications
Shandong University of Technology
2023-2024
Xidian University
2013-2024
Guangxi University of Chinese Medicine
2024
Shandong Normal University
2024
SAIC-GM-Wuling (China)
2024
Shandong University of Science and Technology
2009-2023
National University of Defense Technology
2021-2023
Northwestern Polytechnical University
2023
Women's Hospital, School of Medicine, Zhejiang University
2022
Henan University of Urban Construction
2022
Pedestrian detection is a critical task in computer vision; however, mainstream algorithms often struggle to achieve high accuracy complex scenarios, particularly due target occlusion and the presence of small objects. This paper introduces novel pedestrian algorithm, YOLO-ESL, based on YOLOv7 framework. YOLO-ESL integrates ELAN-SA module, designed enhance feature extraction, with LGA which improves fusion. The module optimizes flexibility efficiency object while effectively multi-scale...
Infertility affects about 10-15% couples over the world, among which a large number of cases underlying causes are still unclear. Recent studies suggest that environmental factors may play an important role in these idiopathic infertilities. Arsenic is heavy metal found drinking water world. Its effect on development female reproductive system at environmental-relevant levels largely unknown. To test hypothesis arsenic exposure during juvenile and puberty affect sex maturation development,...
Synthetic aperture radar (SAR) images have been extensively used in earthquake monitoring, resource survey, agricultural forecasting, etc. However, it is a challenge to interpret SAR with severe speckle noise and geometric deformation due the nature of imaging. The translation SAR-to-optical provides new support for interpretation images. Most existing networks, which are based on generative adversarial networks (GANs), vulnerable part information loss during feature reasoning stage, making...
Recurrent neural networks (RNNs), especially long short-term memory (LSTM) RNNs, are effective network for sequential task like speech recognition. Deeper LSTM models perform well on large vocabulary continuous recognition, because of their impressive learning ability. However, it is more difficult to train a deeper network. We introduce training framework with layer-wise and exponential moving average methods models. It competitive that than 7 layers successfully trained Shenma voice search...
Semantic segmentation of high-resolution remote sensing images (HRSI) is significant, yet challenging. Recently, several research works have utilized the self-attention operation to capture global dependencies. HRSI complex scenes and rich details, implementation on a whole image will introduce redundant information interfere with semantic segmentation. The detail recovery another challenging aspect Several networks use up-sampling, skip-connections, parallel structure, enhanced edge...
Recently, multirotor UAVs have been widely used in high-precision terrain mapping, cadastral surveys and other fields due to their low cost, flexibility, high efficiency. Indirect georeferencing of ground control points (GCPs) is often required obtain highly accurate topographic products such as orthoimages digital surface models. However, practical projects, GCPs are susceptible anomalies caused by external factors (GCPs covered foreign objects crops cars, vandalism, etc.), resulting a...
Genetic algorithm has been widely used in the various optimal problems. Its application fuzzy control is still limited by factors such as local and premature convergence. Therefore, this paper proposes that rule were adjusted together using hybrid based on genetic ant colony algorithm. generates initial candidate develop pheromone of Updating pheromone, operation replaces selecting obtains better candidate. Then, carry out subsequence The simulation results show make controller get...
In this work, the difference convex function based on capped nuclear norm minimization (CNNM-DC) method, a cross-spectral matrix (CSM) completion iterative algorithm with excellent noise immunity, is proposed to realize acoustic imaging of non-synchronous measurements (NSM) under condition low signal-to-noise ratio (SNR). Compared CSM (NNM) model, truncated regularization (TNNR) and weighted (WNNM) method can obtain image narrow main lobes no sidelobe SNR. The simulation results show that,...
Recent work has demonstrated the superiority of compressive sensing (CS) for sound source localization. Using microphone array, direction-of-arrival (DOA) estimation is obtained by using wideband extensions CS technique. For tracking moving sources, a method combined with particle filter (PF) algorithm proposed. Each approximated set particles different weights assigned. Based on DOA estimation, kernel density estimator as likelihood function to estimate probability position. The uniform...
Abstract In this paper, a novel variational model with box constraints is proposed to restore images corrupted by impulse noise. The composed of fractional‐order total variation regularization and L p ‐fidelity term . Moreover, the new possesses advantages preserving sharp edges removing blocking effect. To solve model, some auxiliary variables are first introduced transform it into easy‐to‐solve subproblems. Further, alternating direction method multipliers, iteratively re‐weighted ℓ 1...
As the main raw material in glass manufacturing, quartz sand and its elementary contents play an important role quality. contains a variety of impurities such as metal oxides, which makes traditional chemical analysis methods for determining composition challenging cumbersome, results are often influenced by skill level analysts various reagent factors. In this study, we investigate feasibility employing calibration-free laser-induced breakdown spectroscopy (CF-LIBS) simultaneous...
Purpose To present a HOmologous Black-Bright-blood and flexible Interleaved imaging (HOBBI) sequence for dynamic contrast-enhanced magnetic resonance (MRI) of the vessel wall. Theory Methods A HOBBI is proposed to acquire high-spatial-resolution black-blood high-temporal-resolution bright-blood images in an interleaved fashion. Black-blood allows thin wall evaluation, whereas obtains arterial input function accurately. simulation was performed assess accuracy pharmacokinetic parameters...
A novel blind digital image watermarking scheme based on Schur factorization in contourlet domain is proposed this paper. Firstly, the original cover decomposed by domain, then adopted to selected symmetric blocks determined lowpass subband where watermark information embedded. Additionally, encrypted a pseudorandom sequence generated piecewise linear map before Experimental results show that comparison with similar schemes singular value decomposition, method more invisible and robust...
High computational cost and interference with sidelobes other mainlobes are the main problems of acoustic imaging algorithms for cameras. An improved algorithm that deeply integrates object detection a beamforming is proposed to solve above problems. The YOLO-Fastest employed obtain area specified equipment reduce algorithm. Experimental results show an camera runs method can sidelobe images in real-time.