A5051404919- Advanced Fiber Laser Technologies
- Photonic Crystal and Fiber Optics
- Advanced Clustering Algorithms Research
- Advanced Image and Video Retrieval Techniques
- Laser-Matter Interactions and Applications
- Solid State Laser Technologies
- Face and Expression Recognition
- Data Management and Algorithms
- Recommender Systems and Techniques
- Anomaly Detection Techniques and Applications
- Structural Health Monitoring Techniques
- Text and Document Classification Technologies
- Advanced Fiber Optic Sensors
- Photonic and Optical Devices
- Service-Oriented Architecture and Web Services
- Algorithms and Data Compression
- Video Surveillance and Tracking Methods
- Topic Modeling
- Web Data Mining and Analysis
- Advanced Computational Techniques and Applications
- Laser Material Processing Techniques
- Advanced Software Engineering Methodologies
- Network Security and Intrusion Detection
- Complex Network Analysis Techniques
- Image Retrieval and Classification Techniques
Shenzhen Technology University
2021-2025
Huaqiao University
2015-2024
Beijing Technology and Business University
2019-2024
Fuzhou University
2023
Soochow University
2019-2022
Soochow University
2020
Jimei University
2020
Xi'an Jiaotong University
2017-2019
Xidian University
2019
Shenzhen University
2015-2018
Large-scale data clustering is an essential key for big problem. However, no current existing approach "optimal" due to high complexity, which remains it a great challenge. In this article, simple but fast approximate DBSCAN, namely, KNN-BLOCK proposed based on two findings: 1) the problem of identifying whether point core or not is, in fact, kNN and 2) has similar density distribution its neighbors, neighbor points are highly possible be same type (core point, border noise). DBSCAN uses...
Nowadays, sensor-cloud systems have received wide attention from both academia and industry. Sensor-cloud system not only improves performances of wireless sensor networks (WSNs), but also combines different functional WSNs together to provide comprehensive services. However, a variety malicious attacks threaten the security, such as integrity, authenticity, availability so on. Traditional available security mechanisms (e.g., cryptography authentication) are still vulnerable. Although there...
A stable noise-like (NL) mode-locked Tm-doped fiber laser (TDFL) relying on a nonlinear optical loop mirror (NOLM) was experimentally presented. Different from the previous NL TDFL with NOLM, entire polarization-maintaining (PM) construction utilized in our cavity, which makes oscillator have better resistance to environmental perturbations. The robust can deliver bound-state pulses pulse envelope tunable ~14.1 ns ~23.6 and maximum energy of ~40.3 nJ at repetition rate ~980.6 kHz. Meanwhile,...
Herein, we presented a high energy noise-like (NL) pulse Tm-doped fiber laser (TDFL) system. Relying on the nonlinear amplifying loop mirror (NALM), stable pulses with coherence spike width of ∼317 fs and envelope ∼4.2 ns were obtained from an all polarization-maintaining fiberized oscillator at central wavelength ∼1946.4 nm 3 dB bandwidth ∼24.9 nm. After amplification in all-fiberized TDF amplifier system, maximum average output power ∼32.8 W ∼10.1 µJ obtained, which represents highest NL...
With the dramatic increase of multi-media data on Internet, cross-modal retrieval has become an important and valuable task in searching systems. The key challenge this is how to build correlation between multi-modal data. Most existing approaches only focus dealing with paired They use pairwise relationship for exploring them. However, practice, unpaired are more common Internet but few methods pay attention To utilize both data, we propose a one-stream framework triplet fusion network...
Herein, the efficiently high‐power pulsed Raman lasers with wavelength above 2.1 μm are experimentally demonstrated relying on stimulated scattering. A tailored noise‐like pulse (NLP) fiber laser system centered at ≈1953 nm a maximum output power of ≈10.9 W is served as pump source. By directly pumping section highly Ge‐doped silica fiber, first (centered at≈2139 nm) and second ≈2353 powers ≈3.8 ≈0.25 obtained, respectively, which represent highest NLP these regions, to best knowledge....
In this paper, we demonstrate a simplified one-to-many scheme for efficient mid-infrared (MIR) parametric conversion. Such is based on continuous wave (CW) single longitudinal mode master oscillator power-amplifier (MOPA) fiber system as the signal source and picosecond pulsed MOPA system, exhibiting multiple modes, pump source. The beams are combined co-coupled into piece of 50-mm long 5% MgO-doped PPLN crystal As high ∼3.82 W average power at central idler wavelength ∼3.4 µm achieved when...
The conventional near-infrared laser transmission welding (LTW) process for joining dissimilar transparent polymers is limited by the need to incorporate optical absorbents, which compromises joint performance and raises biocompatibility concerns. To address these issues, this study proposed a surface modification technique using femtosecond ablation prior process. Experiments involved 520 nm of polymers, followed LTW an 808 laser, with subsequent characterization mechanical property...
Density-based clustering is one of the most popular paradigms existing approaches, approaches this kind, such as DBSCAN, recognize clusters data characterized by a fixed scanning radius. However, some flaws are caused radius, e.g., determination proper radius nontrivial. In order to solve these problems, we revise Meanshift, DPeak, etc. based on two new features, i.e., effective and density heat (DHeat). Generally, name revised algorithms DHeat. The underlying idea assumptions: 1) existence...
The existing rail station passenger flow prediction models are inefficient, due to that most of them use single-source data predict. In this paper, a novel method is proposed based on multi-layer LSTM, which integrates multi-source traffic and multi-techniques (including feature selection Spearman correlation time clustering), improve the performance predicting flow. experimental results show techniques integrated in model helpful, obtains higher accuracy outperforms other methods (e.g....