sEMG • Methodology
Methodology – Models
Model architectures

Methodology – Models

LSTM
Sequential dependencies, optional bidirectional
TCN
Causal & dilated convolutions, residual connections
SNN-only
Event-driven spiking
Hybrid TCN–SNN
Parallel feature combination(fusion)
SpikingTCN
Spikes within convolution blocks

Model list & architecture summary

  • LSTMSequence baseline (bi-LSTM optional)
  • TCNCausal/dilated conv + residuals
  • SNN-onlyLIF layers, event-driven
  • Hybrid TCN–SNNTCN features + SNN firing-rate fusion
  • SpikingTCNLIF inside conv blocks

Methodology - Data & Preprocessing Next