This research project seeks to adapt data analysis methods derived from topology and high-dimensional geometry to electromyogram data for classification purposes.

Specifically, the aim of this project is to precisely describe the topological and geometric properties of a high-dimensional manifold sampled by an array of surface Electromyography ("sEMG") sensors as a subject performs a series of predetermined movements.

The data for this project was originally gathered through the study Latent Factors Limiting the Performance of sEMG-Interfaces by Lobov S., Krilova N. et al. published in Sensors. 2018;18(4):1122. doi: 10.3390/s18041122

Source data can be found here: https://archive.ics.uci.edu/ml/datasets/EMG+data+for+gestures

To Do List:

1. Create data set of persistence diagram vectors X done

2. Add random Gaussian noise and recalculate persistence X done - null result

3. normalize persistence images for log reg X done

4. use persim to ID influential image components for filter X done

5. SNF X done

6. Complete persistence filter development - in progress

7. Compare fused similarity templates across gesture classes and within and across subjects

8. Compare fused similarity template variance (static vs clan images) to variance in modalities

   • How does variance within and between modalities effect SNF template outcome

9. Scattering transform