Ruixiang Kan

Mr. Ruixiang Kan

School of Information and Communication, Guilin University of Electronic Technology, Guilin, Guangxi, China
orcid

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
2 December 2023 in Applied Sciences
Verified
Journal Article
Open access
PDF file
MWMei Wang
Mei Wang
QMQingshan Mei
Qingshan Mei
XSXiyu Song
Xiyu Song
XLXin Liu
Xin Liu
RKRuixiang Kan
Ruixiang Kan
HQHongbing QIU
Hongbing QIU
Unsupervised anomalous sound detection by machines holds significant importance within the realm of industrial automation. Currently, the task of machine-based anomalous sound detection in complex industrial settings is faced with issues such as the challenge of extracting acoustic feature information and an insufficient feature extraction capability within the detection network. To address these challenges, this study proposes a machine anomalous sound detection method using the lMS spectrogram and ES-MobileNetV3 network. Firstly, the log-Mel spectrogram feature and the SincNet spectrogram feature are extracted from the raw wave, and the new lMS spectrogram is formed after fusion, serving as network input features. Subsequently, based on the MobileNetV3 network, an improved detection network, ES-MobileNetV3, is proposed in this paper. This network incorporates the Efficient Channel Attention module and the SoftPool method, which collectively reduces the loss of feature information and enhances the feature extraction capability of the detection network. Finally, experiments are conducted on the dataset provided by DCASE 2020 Task 2. Our proposed method attained an averaged area under the receiver operating characteristic curve (AUC) of 96.67% and an averaged partial AUC (pAUC) of 92.38%, demonstrating superior detection performance compared to other advanced methods.

9 May 2023 in Symmetry
Verified
Journal Article
Open access
PDF file
MHMengxiang Huang
Mengxiang Huang
MWMei Wang
Mei Wang
XLXin Liu
Xin Liu
RKRuixiang Kan
Ruixiang Kan
HQHongbing QIU
Hongbing QIU
Environmental sound classification (ESC) tasks are attracting more and more attention. Due to the complexity of the scene and personnel mobility, there are some difficulties in understanding and generating environmental sound models for ESC tasks. To address these key issues, this paper proposes an audio classification framework based on L-mHP features and the SE-ResNet50 model and improves a dual-channel data enhancement scheme based on a symmetric structure for model training. Firstly, this paper proposes the L-mHP feature to characterize environmental sound. The L-mHP feature is a three-channel feature consisting of a Log-Mel spectrogram, a harmonic spectrogram, and a percussive spectrogram. The harmonic spectrogram and percussive spectrogram can be obtained by harmonic percussive source separation (HPSS) of a Log-Mel spectrogram. Then, an improved audio classification model SE-ResNet50 is proposed based on the ResNet-50 model. In this paper, a dual-channel data enhancement scheme based on a symmetric structure is promoted, which not only makes the audio variants more diversified, but also makes the model focus on learning the time–frequency mode in the acoustic features during the training process, so as to improve the generalization performance of the model. Finally, the audio classification experiment of the framework is carried out on public datasets. An experimental accuracy of 94.92%, 99.67%, and 90.75% was obtained on ESC-50, ESC-10 and UrbanSound8K datasest, respectively. In order to simulate the classification performance of the framework in the actual environment, the framework was also evaluated on a self-made sound dataset with different signal-to-noise ratios. The experimental results show that the proposed audio classification framework has good robustness and feasibility.

9 May 2022 in Wireless Communications and Mobile Computing
Verified
Journal Article
Open access
PDF file
RKRuixiang Kan
Ruixiang Kan
MWMei Wang
Mei Wang
zou zhouzou zhou
zou zhou
PZPeng Zhang
Peng Zhang
HQHongbing QIU
Hongbing QIU
The demand for an indoor localization system is increasing, and related research is also becoming more universal. Previous works on indoor localization systems mainly focus on the acoustic signals in Line of Sight (LOS) scenario to obtain accurate localization information, but their effectiveness in Nonline of Sight (NLOS) scenario remains comparatively untouched. These works are usually less efficient as the acoustic signals often bring diffraction, refraction, scattering, energy decays, and so on in NLOS environments. So the system needs adjusting accordingly in a complex NLOS scenario based on NLOS identification results. Therefore, the identification of NLOS acoustic signal turns out to be significant in the indoor localization system. If the system only uses original support vector machine (SVM) to complete NLOS identification, the result turns out to be poor by our test. To address this challenge, we propose a novel indoor localization system, named ZKLocPro, which utilizes an advanced swarm intelligence method to optimize the traditional SVM classification model to deal with NLOS acoustic signal identification. Its results can help the system adjust the localization process if necessary in a complex NLOS scenario. Obviously, it is also significant to build our own NLOS data set, which is suitable for an indoor localization system’s situation. Specifically, four methods are added: (1) new LOS and NLOS acoustic localization signal sample production, rearrangement, and reselecting process; (2) advanced parameter optimization process; (3) elitist strategy; and (4) inertia weight nonlinear decrement. The experimental result shows that our system is efficient and performs better than state-of-the-art congeneric works even in a complex NLOS scenario.

21 March 2023 in Applied Sciences
Verified
Journal Article
Open access
PDF file
RKRuixiang Kan
Ruixiang Kan
MWMei Wang
Mei Wang
XLXin Liu
Xin Liu
XLXiaojuan Liu
Xiaojuan Liu
HQHongbing QIU
Hongbing QIU
For indoor sensor systems, it is essential to implement an extra supporting area notification part. To inform the real-time coordinates, the time difference of arrival (TDOA) algorithm can be introduced. For these indoor localization systems, their main processes are often built based on the line of sight (LOS) scenario. However, obstacles make the off-the-shelf localization system unable to play its due role in the flexible non-line of sight (NLOS) scenario. So, it is necessary to adjust the signals according to the NLOS identification results. However, the NLOS identification methods before were not effective enough. To address these challenges, on the one hand, this paper proposes an adaptive strategy for a dual-receiving signal processing method. On the other hand, the system is matched with the homologous NLOS identification method based on a novel artificial fish school algorithm (AFSA) and the decision tree model. According to our experiments, our novel AFSA optimization method can obtain a better effect and take less time. The NLOS acoustic signal identification accuracy can be improved significantly in flexible scenarios compared with other methods. Based on these processes, the system will achieve more accurate localization results in flexible NLOS situations.

24 October 2024 in Applied Sciences
Verified
Journal Article
Open access
PDF file
XCXu Chen
Xu Chen
MWMei Wang
Mei Wang
RKRuixiang Kan
Ruixiang Kan
HQHongbing QIU
Hongbing QIU
In urban environments, noise significantly impacts daily life and presents challenges for Environmental Sound Classification (ESC). The structural influence of urban noise on audio signals complicates feature extraction and audio classification for environmental sound classification methods. To address these challenges, this paper proposes a Contrastive Learning-based Audio Spectrogram Transformer (CL-Transformer) that incorporates a Patch-Mix mechanism and adaptive contrastive learning strategies while simultaneously improving and utilizing adaptive data augmentation techniques for model training. Firstly, a combination of data augmentation techniques is introduced to enrich environmental sounds. Then, the Patch-Mix feature fusion scheme randomly mixes patches of the enhanced and noisy spectrograms during the Transformer’s patch embedding. Furthermore, a novel contrastive learning scheme is introduced to quantify loss and improve model performance, synergizing well with the Transformer model. Finally, experiments on the ESC-50 and UrbanSound8K public datasets achieved accuracies of 97.75% and 92.95%, respectively. To simulate the impact of noise in real urban environments, the model is evaluated using the UrbanSound8K dataset with added background noise at different signal-to-noise ratios (SNR). Experimental results demonstrate that the proposed framework performs well in noisy environments.

2 November 2023 in Sensors
Verified
Journal Article
Open access
PDF file
RKRuixiang Kan
Ruixiang Kan
HQHongbing QIU
Hongbing QIU
XLXin Liu
Xin Liu
PZPeng Zhang
Peng Zhang
YWYan Wang
Yan Wang
MWMei Wang
Mei Wang
Indoor human action recognition, essential across various applications, faces significant challenges such as orientation constraints and identification limitations, particularly in systems reliant on non-contact devices. Self-occlusions and non-line of sight (NLOS) situations are important representatives among them. To address these challenges, this paper presents a novel system utilizing dual Kinect V2, enhanced by an advanced Transmission Control Protocol (TCP) and sophisticated ensemble learning techniques, tailor-made to handle self-occlusions and NLOS situations. Our main works are as follows: (1) a data-adaptive adjustment mechanism, anchored on localization outcomes, to mitigate self-occlusion in dynamic orientations; (2) the adoption of sophisticated ensemble learning techniques, including a Chirp acoustic signal identification method, based on an optimized fuzzy c-means-AdaBoost algorithm, for improving positioning accuracy in NLOS contexts; and (3) an amalgamation of the Random Forest model and bat algorithm, providing innovative action identification strategies for intricate scenarios. We conduct extensive experiments, and our results show that the proposed system augments human action recognition precision by a substantial 30.25%, surpassing the benchmarks set by current state-of-the-art works.