Improving small-sample performance in portable LIBS analysis with transfer learning and synthetic data-driven CNN models

Key Points

• Regression accuracy significantly improved through the use of transfer learning and synthetic data.
• The lightweight 1D-CNN model satisfies constraints for portable applications and analyzes spectral data.
• Transfer of a model, initially trained on simulated data, enhances performance for practical use.
• This approach may enable more accurate portable analysis, although external validation is needed.