Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Inter-Subject Variance Transfer Learning for EMG Pattern Classification Based on Bayesian Inference

Seitaro Yoneda, Akira Furui

TL;DR

This approach transfers variance information, acquired through pre-training on multiple source subjects, to a target subject within a Bayesian updating framework, thereby allowing accurate classification using limited target calibration data.

Abstract

In electromyogram (EMG)-based motion recognition, a subject-specific classifier is typically trained with sufficient labeled data. However, this process demands extensive data collection over extended periods, burdening the subject. To address this, utilizing information from pre-training on multiple subjects for the training of the target subject could be beneficial. This paper proposes an inter-subject variance transfer learning method based on a Bayesian approach. This method is founded on the simple hypothesis that while the means of EMG features vary greatly across subjects, their variances may exhibit similar patterns. Our approach transfers variance information, acquired through pre-training on multiple source subjects, to a target subject within a Bayesian updating framework, thereby allowing accurate classification using limited target calibration data. A coefficient was also introduced to adjust the amount of information transferred for efficient transfer learning. Experimental evaluations using two EMG datasets demonstrated the effectiveness of our variance transfer strategy and its superiority compared to existing methods.

Inter-Subject Variance Transfer Learning for EMG Pattern Classification Based on Bayesian Inference

TL;DR

This approach transfers variance information, acquired through pre-training on multiple source subjects, to a target subject within a Bayesian updating framework, thereby allowing accurate classification using limited target calibration data.

Abstract

In electromyogram (EMG)-based motion recognition, a subject-specific classifier is typically trained with sufficient labeled data. However, this process demands extensive data collection over extended periods, burdening the subject. To address this, utilizing information from pre-training on multiple subjects for the training of the target subject could be beneficial. This paper proposes an inter-subject variance transfer learning method based on a Bayesian approach. This method is founded on the simple hypothesis that while the means of EMG features vary greatly across subjects, their variances may exhibit similar patterns. Our approach transfers variance information, acquired through pre-training on multiple source subjects, to a target subject within a Bayesian updating framework, thereby allowing accurate classification using limited target calibration data. A coefficient was also introduced to adjust the amount of information transferred for efficient transfer learning. Experimental evaluations using two EMG datasets demonstrated the effectiveness of our variance transfer strategy and its superiority compared to existing methods.

Paper Structure

This paper contains 13 sections, 13 equations, 3 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Proposed method
  3. Problem Setting
  4. Gaussian Classification Model (GCM)
  5. Inter-Subject Transfer Learning
  6. Pre-training with source subjects
  7. Transfer learning to target subject
  8. Motion class prediction
  9. Experiments
  10. Methods
  11. Results
  12. Discussion
  13. Conclusion

Figures (3)

  • Figure 1: Overview of the proposed method
  • Figure 2: Graphical model of GCM for inter-subject classification. The gray and white nodes represent the observed and unobserved variables, respectively.
  • Figure 3: Average classification accuracy with changing $r$ for each dataset. Note that the maximum value of $r$ varies across datasets due to differences in the number of source subjects and data length of them. Each point and the shaded areas represent average accuracy and $95\%$ confidence interval respectively, for all participants. Stars indicate an accuracy corresponding to $r=1$.