Zhiyong Xu

Prof. Dr. Zhiyong Xu

orcid

Research Focus & Metrics

Academic Career Length

16Years (2011-Present)

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

Representative Publications

View All Publications
1 February 2015 in IET Signal Processing
Verified
Journal Article
Open access
PDF file
KQKun Qian
Kun Qian
ZXZhiyong Xu
Zhiyong Xu
HXHuijie Xu
Huijie Xu
YWYaqi Wu
Yaqi Wu
Zhao ZhaoZhao Zhao
Zhao Zhao
Snore related signals (SRS) have been found to carry important information about the snore source and obstruction site in the upper airway of an Obstructive Sleep Apnea/Hypopnea Syndrome (OSAHS) patient. An overnight audio recording of an individual subject is the preliminary and essential material for further study and diagnosis. Automatic detection, segmentation and classification of SRS from overnight audio recordings are significant in establishing a personal health database and in researching the area on a large scale. In this study, the authors focused on how to implement this intelligent method by combining acoustic signal processing with machine learning techniques. The authors proposed a systematic solution includes SRS events detection, classifier training, automatic segmentation and classification. An overnight audio recording of a severe OSAHS patient is taken as an example to demonstrate the feasibility of their method. Both the experimental data testing and subjective testing of 25 volunteers (17 males and 8 females) demonstrated that their method could be effective in automatic detection, segmentation and classification of the SRS from original audio recordings.

16 March 2017 in Acta Oto-Laryngologica
Verified
Journal Article
Closed access
HPHao Peng
Hao Peng
HXHuijie Xu
Huijie Xu
ZXZhiyong Xu
Zhiyong Xu
WHWeining Huang
Weining Huang
RJRuifang Jia
Ruifang Jia
WHWeihong Huang
Weihong Huang
Objective: To discuss the possibility of fundamental frequency (F0) and formant frequency (FF) to generally differentiate the sources of snoring sounds determined by drug-induced sleep endoscopy (DISE). Methods: A total of 74 snoring subjects underwent DISE and snoring sounds were recorded simultaneously. The noise-suppressed snoring sounds were analyzed and classified into different groups based on the sources of vibration identified by DISE. F0 and FFs were calculated. Results: Totally, 516 snoring sounds from three vibrating sources (the palate, combined the palate and the lateral wall, the lateral wall) of 47 patients were divided into three groups then analyzed. The levels of F0 and FFs for each group follow the order: Group 1 < Group 2 < Group 3. There was statistical difference between Group 1 and other groups in F0 and F2 (p < .05). The area under the receiver–operator curves (AUC) was F0, at 0.727, and the cut-off value was 134.2 Hz; and F2, at 0.654, and the cut-off value was 2028.0 Hz. Conclusions: F0 and the second formant frequency (F2) are found to be significantly lower in palatal snoring sound. F0 might be a significant in distinguishing palatal snoring sound from non-palatal snoring sound. F2 is more significant than F1 and F3 in identifying the sources of the snoring sounds but is less sensitive than F0.

24 March 2023 in IEEE Transactions on Geoscience and Remote Sensing
Verified
Journal Article
Closed access
ZXZhiyong Xu
Zhiyong Xu
XFXiaolin Feng
Xiaolin Feng
STSirui Tian
Sirui Tian
XSXiang-Jun Shen
Xiang-Jun Shen
HZHong Zhang
Hong Zhang
CWChao Wang
Chao Wang
Synthetic aperture radar (SAR) image despeckling is a challenging task as speckle noise is spatially correlated and signal-dependent, and appears as a grainy texture superimposed on images. Although traditional low-rank SAR image despeckling methods have shown promising performance, they have the problem of producing over-smoothed images with blurred edges due to their low-rank characteristics. In this paper, we propose a novel edge preserved SAR despeckling method named EP-LRSID, which can keep rich edge details while reducing speckle noise. Specifically, EP-LRSID takes a fresh look at the low-rank model, i.e., we can obtain structural edge information from residuals which is viewed as noise and simply disregarded by the traditional low-rank methods. To obtain discriminative edge information from residuals, the edge subspace is obtained in a manifold framework by using a dynamic affinity graph regularization. Moreover, a new hierarchical prior knowledge regulation is designed to make different kinds of pixels processed hierarchically, especially the strong scattering points in SAR images. By introducing this prior knowledge, our low-rank model can obtain more confidential low-rank parts and edge parts, thus structural information including edges can be better preserved in this way. Extensive experiments on several real and synthetic datasets demonstrate that EP-LRSID can achieve the highest despeckling performance with edge preservation than other state-of-the-art despeckling algorithms.

27 April 2023 in Remote Sensing
Verified
Journal Article
Open access
PDF file
XFXiaolin Feng
Xiaolin Feng
STSirui Tian
Sirui Tian
SAStanley Ebhohimhen Abhadiomhen
Stanley Ebhohimhen Abhadiomhen
ZXZhiyong Xu
Zhiyong Xu
XSXiang-Jun Shen
Xiang-Jun Shen
CWChao Wang
Chao Wang
The low-rank models have gained remarkable performance in the field of remote sensing image denoising. Nonetheless, the existing low-rank-based methods view residues as noise and simply discard them. This causes denoised results to lose many important details, especially the edges. In this paper, we propose a new denoising method named EPLRR-RSID, which focuses on edge preservation to improve the image quality of the details. Specifically, we considered the low-rank residues as a combination of useful edges and noisy components. In order to better learn the edge information from the low-rank representation (LRR), we designed multi-level knowledge to further distinguish the edge part and the noise part from the residues. Furthermore, a manifold learning framework was introduced in our proposed model to better obtain the edge information, as it can find the structural similarity of the edge part while suppressing the influence of the non-structural noise part. In this way, not only the low-rank part is better learned, but also the edge part is precisely preserved. Extensive experiments on synthetic and several real remote sensing datasets showed that EPLRR-RSID has superior advantages over the compared state-of-the-art (SOTA) approaches, with the mean edge protect index (MEPI) values reaching at least 0.9 and the best values in the no-reference index BRISQUE, which represents that our method improved the image quality by edge preserving.

7 April 2023 in IEEE Sensors Journal
Verified
Journal Article
Closed access
Zhao ZhaoZhao Zhao
Zhao Zhao
HKHongrui Kan
Hongrui Kan
JLJiale Lin
Jiale Lin
ZXZhiyong Xu
Zhiyong Xu
Direction of arrival (DOA) estimation is the key to many audio applications. Recently, sparse component analysis (SCA)-based methods have attracted much attention, in which single-source points (SSPs) and single-source zones (SSZs) where one source is dominant over the others in time-frequency domain are usually detected to construct the pooled histogram containing multi-source DOA information. Nonetheless, the SSZ size in existing methods is fixed and empirically predetermined, which cannot accommodate to the varying spectro-temporal property of speech sources. Furthermore, higher SSP concentration in a SSZ implies a locally stronger dominant source as well as more reliable DOA information extracted therein, which however is also not taken into account yet. To address these problems, a DOA estimation algorithm for multiple speech sources based on flexible SSZs and concentration weighting is presented in this paper. First, in each frame, correlation coefficients of time delay vectors across adjacent frequency bins are calculated to identify SSPs, followed by flexible SSZs construction using varying number of SSPs located at consecutive frequency bins. Next, the number of SSPs in each flexible SSZ is considered as a proxy of corresponding concentration degree, and employed as weighting factor to form the pooled histogram. Finally, a matching pursuit (MP)-based approach is utilized to obtain multi-source DOA estimates. Simulation results reveal that the proposed method significantly outperforms existing approaches in terms of noise floor in pooled histogram, angular resolution, and performance under various signal-to-noise ratio and reverberant conditions. Real-world experiments also verify its effectiveness, and meanwhile demonstrate considerably reduced computational complexity.

17 December 2024 in IEEE Transactions on Audio, Speech and Language Processing
Verified
Journal Article
Closed access
SLShiqin Li
Shiqin Li
Zhao ZhaoZhao Zhao
Zhao Zhao
ZXZhiyong Xu
Zhiyong Xu
YRYingjiao Rong
Yingjiao Rong
With booming growth of Internet of Things technique, acoustic sensor networks comprised of microphone array nodes have been increasingly applied to multi-source speech enhancement (SE). Compared with conventional centralized SE algorithms, distributed approaches achieve comparable performance with many advantages such as convenience, low cost and scalability. However, existing distributed methods either need to run once for each desired source independently, or require the number of channels transmitted by each node to increase proportionally with the number of desired sources. Furthermore, it is also challenging for each node to acquire all the steering vectors as demanded by existing methods. To address these issues, a node-specific distributed generalized sidelobe canceler (NS-DGSC) algorithm is proposed in this article. Here, “node-specific” refers to scenarios where each node estimates individual desired source. First, microphone signals at each node are pre-filtered by a local generalized sidelobe canceler (GSC), yielding preliminary enhancement for individual desired source. Then, for each node, preliminary results exchanged from other nodes are utilized to expand the lower branch of local GSC. Finally, the expanded lower branch is employed to calculate a new noise canceler, leading to the final output at each node. Compared with state-of-the-art distributed algorithms, our approach only requires single-channel transmission at each node as well as available local steering vectors for simultaneously enhancing multiple desired sources. Simulations across various signal-to-interference-plus-noise ratio (SINR) input, reverberation time, source-to-node distance, and steering vector estimation error conditions reveal that the proposed approach considerably outperforms existing methods in terms of signal-to-distortion ratio (SDR) and robustness in practice, while provides comparable SINR improvement. Real-world experiments also verify its effectiveness.