XIN GUO

Mr. XIN GUO

orcid

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

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

Representative Publications

View All Publications
7 January 2025 in Sensors
Verified
Journal Article
Open access
PDF file
XGXIN GUO
XIN GUO
Young KimYoung Kim
Young Kim
XNXueli Ning
Xueli Ning
Se Dong MinSe Dong Min
Se Dong Min
The Transformer model has received significant attention in Human Activity Recognition (HAR) due to its self-attention mechanism that captures long dependencies in time series. However, for Inertial Measurement Unit (IMU) sensor time-series signals, the Transformer model does not effectively utilize the a priori information of strong complex temporal correlations. Therefore, we proposed using multi-layer convolutional layers as a Convolutional Feature Extractor Block (CFEB). CFEB enables the Transformer model to leverage both local and global time series features for activity classification. Meanwhile, the absolute position embedding (APE) in existing Transformer models cannot accurately represent the distance relationship between individuals at different time points. To further explore positional correlations in temporal signals, this paper introduces the Vector-based Relative Position Embedding (vRPE), aiming to provide more relative temporal position information within sensor signals for the Transformer model. Combining these innovations, we conduct extensive experiments on three HAR benchmark datasets: KU-HAR, UniMiB SHAR, and USC-HAD. Experimental results demonstrate that our proposed enhancement scheme substantially elevates the performance of the Transformer model in HAR.

7 June 2024 in IEEE Access
Verified
Journal Article
Open access
XNXueli Ning
Xueli Ning
Young KimYoung Kim
Young Kim
Se Dong MinSe Dong Min
Se Dong Min
XGXIN GUO
XIN GUO
JHJong Gab Ho
Jong Gab Ho
Balance training is widely used to improve stability, and Achilles tendon vibration is an effective method. However, evaluation of training progress often relies on Center of Pressure (COP) analysis, which can be challenging for non-experts. To provide an objective and automated assessment, this study explores machine learning techniques. Achilles tendon vibration was applied during standing, and COP data were collected under various conditions, including eyes open/closed and cognitive/non-cognitive tasks. To more accurately assess the training effects, this study applied machine learning techniques that combine wavelet decomposition for feature extraction. Three genetic algorithm-based machine learning models (GA-SVM, GA-LGBM, and GA-LR) were constructed for feature selection and classification. The results showed that all three models achieved classification accuracies above 80% in identifying Achilles tendon vibration and non-vibration data, with SVM achieving the highest accuracy of 89.59%. Among the selected features, entropy category features played a crucial role, and entropy values were higher under Achilles tendon vibration conditions than under non-vibration conditions. This study confirms the feasibility of applying machine learning to Achilles tendon vibration rehabilitation training in the future, and the identified key features also provide a theoretical basis for the analysis of Achilles tendon vibration data. These findings provide valuable insights for further optimization of balance rehabilitation training programs.