Improving small-sample performance in portable LIBS analysis with transfer learning and synthetic data-driven CNN models
Abstract
A lightweight 1D-CNN designed for portable devices was developed. By transferring a model trained on simulated data to portable spectral data, regression accuracy was improved while satisfying portable application constraints.
Fields of Study & Subjects
Laser-induced spectroscopy and plasma
Mass Spectrometry Techniques and Applications
Spectroscopy Techniques in Biomedical and Chemical Research
Publication details
- Journal / source
- Journal of Analytical Atomic Spectrometry
- Publisher
- Royal Society of Chemistry (RSC)
- Volume
- 41
- Issue
- 2
- Pages
- 655–665
- Publication date
- 1 January 2026
- Language
- EN