Preprocessing Pipeline and Application of CNNs for Surface Electromyography (sEMG) Signal Classification

Year: Jun – Dec 2025

sEMGCNNResNetAlexNetSignalClassificationDRDO

Institute of Nuclear Medicine and Allied Sciences-DRDO, Ministry of Defence · Machine Learning Research Intern

  • Achieved 98% accuracy by developing a ResNet, CNNs to classify surface electromyography (sEMG) stress measurements.
  • Improved classification performance from 91% to 97.98% through architecture redesign and training, showing raw sEMG inputs outperform PyEMGPipeline preprocessing.
  • Implemented AlexNet alongside ResNet experiments to further refine accuracy and generalization across sEMG datasets.
  • Selected to present this research on stress-level classification at the SUMMIT 2.0 Conference.

Resources & Links

Continue Exploring

Discover other related projects or articles.

Back to All Projects