JMJiaqing Mo

Jiaqing Mo

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Jiaqing Mo has a total of 54 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Recognition algorithm for optic fiber vibration signals based on improved CLDNN

    by Yiteng Cui, Jiaqing Mo, Jinshui Qi

    Current optical fiber vibration signal (OFVS) recognition algorithms are limited by insufficient convolutional layer receptive fields, typically relying on a single convolutional layer for feature extraction. This approach often overlooks critical long-range dependencies between channels and fails to capture fine-grained features, leading to difficulties in distinguishing fiber optic vibration signals with similar energy levels and waveforms. To address these shortcomings, this paper proposes the FDSF-CLDNN model (Fusing Dependencies and Small Local Features), building on the CLDNN framework. The GCSconv module is introduced to emphasize long-range dependencies and local fine features, thereby enhancing the extraction of both global channel and spatial features, allowing the model to focus more effectively on key features and spatial locations. Additionally, to preserve fine-grained features while maintaining long-range channel dependencies, this paper proposes the EVC-SCconv module. This module integrates a self-calibration mechanism to amplify inter-channel information, extend the convolutional kernel's receptive field, and accentuate the salient characteristics of the vibration signals. Simultaneously, it leverages an attention mechanism to retain global feature dependencies while highlighting local small features. Furthermore, to maintain time-domain signal dependencies while reducing computational complexity and improving model efficiency, the LSTM module in the original architecture is replaced by the GRU module, which achieves computational savings without sacrificing recognition accuracy. Experimental results demonstrate that the proposed FDSF-CLDNN model achieves a recognition accuracy of 99.0% for optical fiber vibration signals, outperforming other models in terms of both accuracy and stability. Moreover, the proposed model converges more rapidly and yields higher average recognition accuracy compared to existing approaches.

  • Open Access

    WDE-TransUNet: Brain Tumor Segmentation in MRI Images Utilizing Detail Enhancement and Transformers

    by Xinhu Ni, Jiaqing Mo

    WDE-TransUNet is a novel neural network architecture specifically designed for robust and precise semantic segmentation in medical imaging. This method addresses the challenges posed by inconsistent image quality, diverse pathologies, and anatomical variations due to different imaging techniques by integrating Transformer and convolutional neural network (CNN) modules within an encoder-decoder framework. It utilizes prior knowledge to assign weights at the input stage, optimizing feature utilization across modalities and mitigating the impact of irrelevant features. The encoder employs detail enhanced convolution (DEConv) and root mean square (RMS)Transformer modules to enhance local feature extraction, and then conveys these features to the decoder via skip connections, enriching both spatial and channel information. Evaluated on the Brain Tumor Segmentation Challenge (BraTS)2019, 2020, and 2021 datasets, this method achieved outstanding performance, with average Dice scores of 84.74%, 86.64%, and 90.66%, respectively. Our experimental results demonstrate that WDE-TransUNet, by integrating RSM Transformer with CNN in an encoder-decoder configuration, is superior to existing algorithms for brain tumor segmentation.


  • Recognition of Fiber Optic Vibration Signals Based on Laplace Wavelet Transform and Deep Learning

    by Jinshui Qi, Jiaqing Mo, Yasen Niu, Yiteng Cui

    Convolutional neural networks possess the capability of feature learning and nonlinear mapping, which has significant advantages in classifying and recognizing optical fiber vibration signals. In order to further enhance the recognition rate of vibration signals, this paper combines wavelet transform with convolutional neural networks and designs a convolutional layer based on parameterized wavelets. In this layer, the initial signal is convolved with parameterized Laplace wavelet dictionaries to complete the wavelet transform. Such customized filters make more sense than filters with randomly initialized parameters for traditional CNNs. Simultaneously, we introduce the channel attention mechanism to enhance the features of the filtered signals. Subsequently, standard CNNs are employed to extract and process signal features, ultimately utilizing a multi‐layer perceptron for recognition and classification. The experimental results show that the network model possesses better recognition refinement. © 2024 Institute of Electrical Engineer of Japan and Wiley Periodicals LLC.


  • Open Access

    Multi-objective model predictive control method for T-type three-level grid-connected

    by Ning Pan, Zhen Xie, Jiaqing Mo, Yuanhao Wang

    Based on the mathematical model of T-type three-level topology, a T-type three-level inverter model is proposed to predict the multi-objective control method, which mainly includes DC neutral voltage (NP-V), switching state, inverter voltage vector, and multi-objective cost function. Different emphases can be selected according to the application requirements, and coordinated control of different control objectives can be realized by giving different weight values to different control objectives. An extended observer (ESO) is added to observe the error part of the grid-connected inverter system, and the observed results are compensated to the grid-connected reference voltage. The simulation results show that the total harmonic distortion rate of the output current of the inverter is lower than that of the traditional finite control set model predictive control strategy. In addition, the proposed method can realize the balance control of the midpoint potential, and the principle is simple.

  • GTDNet: A Dual-branch Domain Adaption Network for Graphic-rich Text Detection

    by Tianci Zhang, Gang Zhou, Yajun Liu, Yangxin Liu, En Deng, Jiaqing Mo

    In recent years, text detection methods based on convolutional neural networks have been extensively studied and obtained successful results in various datasets. However, many graphic- rich text instances are still difficult to detect due to their irregular appearance and easy confusion with graphics. To address this issue, we construct a dual-branch domain adaption network for graphic-rich text detection (GTDNet). Due to the coexistence of ordinary text (source domain) and graphic-rich text (target domain) in real scenarios, our dual-branch domain adaptation network can complete two tasks well at the same time. We incorporate a feature alignment module in one branch, and use an attension fusion module for fully combining two branch features. We have created a clothing printing dataset that comprises both ordinary text and graphic-rich text. Ex-tensive experiments on this dataset show that GTDNet achieves best performance on graphic-rich texts with a hardly decline performance on ordinary texts.


  • Open Access

    Class Imbalanced Medical Image Classification Based on Semi-Supervised Federated Learning

    by Wei Liu, Jiaqing Mo, Furu Zhong

    In recent years, the application of federated learning to medical image classification has received much attention and achieved some results in the study of semi-supervised problems, but there are problems such as the lack of thorough study of labeled data, and serious model degradation in the case of small batches in the face of the data category imbalance problem. In this paper, we propose a federated learning method using a combination of regularization constraints and pseudo-label construction, where the federated learning framework consists of a central server and local clients containing only unlabeled data, and labeled data are passed from the central server to each local client to take part in semi-supervised training. We first extracted the class imbalance factors from the labeled data to participate in the training to achieve label constraints, and secondly fused the labeled data with the unlabeled data at the local client to construct augmented samples, looped through to generate pseudo-labels. The purpose of combining these two methods is to select fewer classes with higher probability, thus providing an effective solution to the class imbalance problem and improving the sensitivity of the network to unlabeled data. We experimentally validated our method on a publicly available medical image classification data set consisting of 10,015 images with small batches of data. Our method improved the AUC by 7.35% and the average class sensitivity by 1.34% compared to the state-of-the-art methods, which indicates that our method maintains a strong learning capability even with an unbalanced data set with fewer batches of trained models.


  • Open Access

    Optical Fiber Vibration Signal Identification Method Based on Improved YOLOv4

    by Jiangwei Zhang, Jiaqing Mo, Xinrong Ma, Jincheng Huang, Fubao Song

    In the traditional peripheral-security-early-warning system, the endpoint detection and pattern recognition of the signals generated by the distributed optical fiber vibration sensors is completed step-by-step and in an orderly manner. The method by which these two processes may be placed end-to-end in a network model and processed simultaneously to improve work efficiency has increasingly become the focus of research. In this paper, the target detection algorithm combines the endpoint-detection and pattern-recognition processes of the vibration signal, which can not only quickly locate the start and end vibration positions of the signal but also accurately identify a certain type of signal. You Only Look Once v4 (YOLOv4) is one of the most advanced target detection algorithms, achieving the optimal balance of speed and accuracy. To reduce the complexity of the YOLOv4 model and solve the dataset’s unbalanced sample classification problem, we use a deep separable convolution (DSC) network and a focal loss function to improve the YOLOv4 model. In this paper, the five kinds of signals collected in real-time are visualized as two different datasets in oscillograph and time-frequency diagrams as detection objects. According to the experimental results, we obtained 98.50% and 93.48% mean Average Precision (mAP) and 84.8 and 69.9 frames per second (FPS), respectively, which are improved compared to YOLOv4. Comparing the improved algorithm with other optical fiber vibration signal recognition algorithms, the mAP and FPS values were improved, and the detection speed was about 20 times faster than that of other algorithms. The improved algorithm in this paper can quickly and accurately identify the vibration signal of external intrusion, reduce the false-alarm rate of the early-warning system, and improve the real-time detection rate of the system while ensuring high recognition accuracy.


  • Open Access

    A Fiber Vibration Signal Recognition Method Based on CNN-CBAM-LSTM

    by Jincheng Huang, Jiaqing Mo, Jiangwei Zhang, Xinrong Ma

    By trying to solve the issue of identifying multiple types of intrusion vibration signals collected by distributed vibrating fiber optic sensors, this study investigates the signal identification and feature extraction of intrusion signals, and proposes an optical fiber vibration signal (OFVS) identification method based on deep learning. The external vibration signal is collected by the Sagnac fiber optic interferometer, and then denoised by spectral subtraction. Endpoint detection is carried out by combining the short-time logarithmic energy method and the spectral entropy method. Finally, the equal-length signal containing valid information is intercepted and the corresponding preprocessing is carried out. The method for feature processing incorporates the strong feature learning capability of the long-short-term memory (LSTM) and the great short-term feature extraction capability of the convolutional neural network (CNN). At the same time, to further enhance the signal feature identification, a convolutional block attention module (CBAM) is introduced to perform adaptive feature refinement on the signal. In summary, a network model combining CNN, LSTM, and CBAM is proposed to process the signal features, and finally, the multi-layer perceptron (MLP) is used to complete the task of classification and recognition of multi-type intrusion signals. The experimental findings indicate that the OFVS method of CNN-CBAM-LSTM can effectively identify four kinds of OFVS, and the overall average recognition accuracy reaches 97.9%. Walking and knocking signals among them are recognized with over 99% accuracy.