Yongbiao Wang

Mr. Yongbiao Wang


Research Focus & Metrics

Academic Career Length

5Years (2022-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
11 February 2022 in Applied Sciences
Verified
Journal Article
Open access
PDF file
YWYahui Wang
Yahui Wang
WZWenxi Zhang
Wenxi Zhang
ZWZhou Wu
Zhou Wu
XKXinxin Kong
Xinxin Kong
YWYongbiao Wang
Yongbiao Wang
HZHongxin Zhang
Hongxin Zhang
DNN-based Speech Enhancement (SE) models suffer from significant performance degradation in real recordings due to the mismatch between the synthetic datasets employed for training and real test sets. To solve this problem, we propose a new Generative Adversarial Network framework for Noise Modeling (NM-GAN) that creates realistic paired training sets by imitating real noise distribution. The proposed framework combines a novel 7-layer U-Net with two bidirectional long short-term memory (LSTM) layers that act as a generator to construct complex noise. NM-GAN generates enough recall (diversity) and precision (noise quality) in its samples through adversarial and alternate training, effectively simulating real noise, which is then utilized to compose realistic paired training sets. Extensive experiments employing various qualitative and quantitative evaluation metrics verify the effectiveness of the generated noise samples and training sets, demonstrating our framework’s capabilities.

10 November 2026 in Photonics
Verified
Journal Article
Open access
PDF file
YYYeping Yin
Yeping Yin
ZWZiqi Wei
Ziqi Wei
YWYongbiao Wang
Yongbiao Wang
XKXinxin Kong
Xinxin Kong
ZWZhou Wu
Zhou Wu
WZWenxi Zhang
Wenxi Zhang
Random scattering from a rough target surface causes a random phase distribution in the echo light field of a laser Doppler vibrometer (LDV), which in turn affects the interference between the echo light and the reference light. Based on related theories including Gaussian beam propagation and rough surface modeling, theoretical models for light field propagation, random target scattering, and interference intensity were established. Numerical simulation was conducted for a typical application scenario with a laser wavelength of 1550 nm, and a measurement distance of 1 m was used to investigate the effects of transceiver optical system parameters and wavefront aberrations on system performance. The results indicate that both the F-number of the transceiver lens and the wavefront aberrations have a significant impact on the average interference intensity. Under rough target conditions, to achieve a higher average interference intensity, the designed F-number of the transceiver lens should be approximately 70–85% of the F-number of the focusing lens. Regarding wavefront aberrations, defocus, 0∘ primary astigmatism, primary spherical aberration, and 0∘ secondary astigmatism have a more significant effect on the average interference intensity. For applications involving rough targets, the design tolerance values for the Zernike coefficients of these four aberrations are all higher than those for mirror-finished targets. For these four aberrations, the design tolerance values for ground surfaces are increased by 16.7–66.7%, while the design tolerance values for milled surfaces are increased by 11.1–66.7%, effectively reducing the difficulty of lens design. The results of this study provide a theoretical reference for the design of optical system parameters in LDV and other coherent detection systems.