YGYanli Geng

Yanli Geng

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

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Yanli Geng has a total of 27 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Research on Abnormal Lung Sound Recognition and Diagnosis Based on Improved CNN and Transfomer

    by Yanli Geng, Zelei Li, Hao Sun

    Lung auscultation is one of the effective means to diagnose lung diseases. This study aims at the lung sounds of patients, and studies the diagnosis and recognition method of abnormal lung sounds in patients with lung diseases by combining convolutional neural network CNN (Convolutional Neural Network) and Transformer. In this study, a realtime lung sound acquisition system was designed, and the lung sound information was uploaded to the PC terminal by using the wireless WIFI module. Under the guidance of professional physicians, the collected lung sounds of patients were labeled to establish a feature data set (normal, light, common, and heavy). Secondly, based on the Butterworth filter Gaussian filter and the nonnegative matrix factorization method, the lung sound denoising was performed and the MFCC feature information of the lung sound was extracted. Finally, the lung sound auxiliary diagnosis model was designed to analyze the characteristics of lung sounds, and the patient's condition was diagnosed and evaluated. This paper proposes an SCBAM-CNN-Transformer classification model. After optimizing CNN by improving the channel attention mechanism CBAM, the improved CNN model is combined with Transformer. It combines the advantages of CNN spatial feature representation and Transformer sequence coding, forming a strong global feature perception and local feature extraction ability. The algorithm proposed in this paper has good robustness and generalization ability, and the recognition accuracy is 94.8 %. The experimental results show that the proposed algorithm can more effectively identify abnormal lung sounds and determine the infection level than other deep learning models.


  • Pain Expression Recognition Method Based on Fusion Attention Mechanism Capsule Network

    by Yanli Geng, Li Yin, Shijie Lu, Sun Hao

    Facial expression serves as a crucial indicator for pain recognition. However, existing approaches often face challenges in accurately extracting features from complex spatial information in facial expressions. To address this, this study introduces a pain recognition method based on a fusion attention mechanism capsule network. The main contributions of this paper are outlined as follows. Firstly, a comprehensive pain expression database is established by using video image sequences. Subsequently, the capsule attention mechanism is introduced after the primary capsule layer of the network, facilitating more effective feature extraction and improving the accuracy of information transmission within the capsule network. Additionally, the fully connected decoder in the capsule network is replaced by the deconvolution decoder, resulting in a reduction in the overall parameter count. Experimental results on the pain expression database demonstrate the efficacy of the proposed capsule network, achieving an impressive accuracy of 98.21%. Importantly, the proposed approach outperforms other deep learning models in terms of accuracy for pain recognition.

  • RESEARCH ON PIG BODY SIZE MEASUREMENT SYSTEM BASED ON STEREO VISION

    by Yanli Geng, Xiaodong Yue, Yankai Ji, Yanbo Lin, Yanfang Fu, Shucai Yang

    Body size of pigs is an important evaluation indicator in pig breeding. The traditional method of body size measurement is usually in manual way, which requires more employees and causes stress reactions of pigs. In response to the shortcomings of the traditional methods, this paper designed a system for measuring the body size of pigs based on stereo vision. The point cloud of both the calibration object and the pig was collected using dual KinectV2 cameras. Pre-processing was conducted using filtering and random sample consensus to remove background noise from the point clouds. As there was limited overlap between the two sides of the point clouds, the rotation matrix obtained from registering the calibration object was applied to the pig point clouds. Curve fitting and slicing were then utilized to measure the pig's body dimensions, including length, width, height, and abdominal circumference. The results of the study indicated that the mean absolute percentage error (MAPE) was 2.13% for body length, 1.02% for body width, 1.05% for body height, and 2.21% for abdominal girth. These results demonstrate the high accuracy and practical production value of the system.


  • Model‐free robust adaptive integral sliding mode impedance control of knee–ankle–toe active transfemoral prosthesis

    by Zhengen Wu, Yawei Chen, Yanli Geng, Xirui Wang, Bokai Xuan

    Background: Wearing appropriate active prosthesis is the guarantee of daily life for amputees. Normally the controller of the traditional active transfemoral prosthesis is designed based on the mathematical model. The modelling error and the external interference will reduce the control accuracy of the system and make the prosthesis unable to operate in the desired trajectory.Methods: Firstly, combined with time delay estimation (TDE), a model‐free robust integral sliding mode impedance controller is designed. This method not only suppress the impedance error, but also eliminate the nonlinear relationship and disturbance in the dynamic model. Secondly, an adaptive law is proposed to update the controller gain, which provide stable control effect. Thirdly, the stability of prosthesis closed‐loop system is proved by Lyapunov stability theory. Finally, the motor torque is used to drive each joint, and Matlab/Simscape is used to verify the prosthesis control system.Results: From the result of the simulation experiment, the control method has a good tracking effect on each joint. The root mean square error and mean absolute errors of each joint's angle tracking error are 0.6123°, 1.9976°, 0.5574° and 0.2635°, 1.8175°, 0.4796°. Compared with the controller without adaptive gain and impedance control, the control effect is improved, and the plantar pressure of amputees is closer to the sound side.Conclusions: Comparing the results of different controllers, the adaptive integral sliding mode impedance controller with TDE can better track the expected angles of each joint. The gait is more normal. The walking performance of the prosthesis wearers is improved.


  • Music Genre Trend Prediction Based on Spatial-Temporal Music Influence and Euclidean Similarity

    by Cuifeng Qin, Haoran Yang, Wanting Liu, Sanbo Ding, Yanli Geng

    This paper is concerned with the music genre trends prediction according to the music influence and similarity through data analysis. By a Deepwalk and Cosine Similarity method, a mathematical model is formulated for the music influence, which evaluates the influence of musicians and between musicians. The music similarity model is proposed based on the Principal Component Analysis and Euclidean distance, which summarizes some phenomena according to genres. The genre-independent influence reflects the information about cultural influence. The visual demonstration of influence and similarity are provided. Finally, the trends of music genre are summarized and predicted from the music influence and similarity.


  • Open Access

    Modeling and Control of Knee-Ankle-Toe Active Transfemoral Prosthesis

    by Yawei Chen, Bokai Xuan, Yanli Geng, Sanbo Ding, Lingling Chen

    The main joints of lower limb are hip, knee, ankle, and toe. Because of accidents or diseases, some people have to amputate. Wearing prosthesis can help them return to normal life. In the past, Active Knee joint and Active Knee-Ankle transfemoral prosthesis are studied. While in the stance phase, toe joint play an important role. In this article, the dynamic model and control method of Knee-Ankle-Toe Active Transfemoral Prosthesis are studied. The main contents are as follows: Firstly, according to the ground reaction force, the gait cycle is divided into three phases, Early and Middle Stance Phase (E&MSP), Late Stance Phase (LSP) and Swing Phase (SP), and the mechanical model of Knee-Ankle-Toe Active Transfemoral Prosthesis (KATATP) is established. Secondly, consider the cooperative movement of amputee and prosthesis, the kinematics model of prosthesis is established by Lie Groups and Lie Algebras, and the dynamic model of prosthesis is established by Lagrange equation. Thirdly, the whole gait cycle is a complete system and each phase is a different subsystem, switching rules and controllers of different subsystems are designed. Finally, experiments show that KATAP with the switching control enables amputee to have a biomechanically appropriate walking gait and achieve smooth transition between different gait phases.


  • Open Access

    Selective Ensemble Based on Extreme Learning Machine for Sensor-Based Human Activity Recognition

    by Yiming Tian, Jie Zhang, Lingling Chen, Yanli Geng, Xitai Wang

    Sensor-based human activity recognition (HAR) has attracted interest both in academic and applied fields, and can be utilized in health-related areas, fitness, sports training, etc. With a view to improving the performance of sensor-based HAR and optimizing the generalizability and diversity of the base classifier of the ensemble system, a novel HAR approach (pairwise diversity measure and glowworm swarm optimization-based selective ensemble learning, DMGSOSEN) that utilizes ensemble learning with differentiated extreme learning machines (ELMs) is proposed in this paper. Firstly, the bootstrap sampling method is utilized to independently train multiple base ELMs which make up the initial base classifier pool. Secondly, the initial pool is pre-pruned by calculating the pairwise diversity measure of each base ELM, which can eliminate similar base ELMs and enhance the performance of HAR system by balancing diversity and accuracy. Then, glowworm swarm optimization (GSO) is utilized to search for the optimal sub-ensemble from the base ELMs after pre-pruning. Finally, majority voting is utilized to combine the results of the selected base ELMs. For the evaluation of our proposed method, we collected a dataset from different locations on the body, including chest, waist, left wrist, left ankle and right arm. The experimental results show that, compared with traditional ensemble algorithms such as Bagging, Adaboost, and other state-of-the-art pruning algorithms, the proposed approach is able to achieve better performance (96.7% accuracy and F1 from wrist) with fewer base classifiers.