Undergraduate Research Project – Data Collection and Labeling for Synthetic Signal Generation
Data Collection
Generated synthetic power trace datasets for side-channel analysis research using real signal data as ground truth. Developed a pipeline to identify points of interest in raw power traces and reconstruct realistic segments for further AI model training. Statistically validated the generated datasets to ensure high intra-group similarity and strong inter-group differentiation.
• Collected and processed signal data from the DPAcontest v2 dataset
• Labeled important segments using signal derivatives and statistical criteria
• Generated synthetic signals through interpolation techniques for AI pipeline input
• Evaluated dataset quality with statistical metrics to ensure labeling effectiveness.
Generated synthetic power trace datasets for side-channel analysis research using real signal data as ground truth. Developed a pipeline to identify points of interest in raw power traces and reconstruct realistic segments for further AI model training. Statistically validated the generated datasets to ensure high intra-group similarity and strong inter-group differentiation. • Collected and processed signal data from the DPAcontest v2 dataset • Labeled important segments using signal derivatives and statistical criteria • Generated synthetic signals through interpolation techniques for AI pipeline input • Evaluated dataset quality with statistical metrics to ensure labeling effectiveness.
2024 - 2025