Template-based compression and QR-code encoding of ¹H NMR spectra for metabolite quantification and geographical origin authentication of Lycium barbarum L

A template-based feature extraction, spectral compression, and QR-code encoding framework was developed to represent proton nuclear magnetic resonance (¹H NMR) data of Lycium barbarum L. (L. barbarum) in quick response (QR) codes. Assigned spectral regions of interest were parameterized into compact feature matrices and subsequently compressed and encoded into QR codes, enabling a unified workflow for portable decoding, spectral reconstruction, metabolite quantification, and geographical origin authentication. The proposed approach was validated using 136 spectra from five producing regions. The overall workflow reduced the original spectral data volume to 0.20% of its initial size in the final QR-ready representation, with successful QR generation for all samples. Quantitative evaluation of 65 metabolites showed excellent agreement with conventional integration methods, achieving median mean absolute percentage error (MAPE) and symmetric mean absolute percentage error (SMAPE) values of 0.44% and 0.44%, respectively, a median Pearson's correlation coefficient of 1.00, and a median R²-like statistic of 1.00. Among six machine learning models applied to the template-derived feature matrices, the optimized multilayer perceptron achieved the highest classification performance, with an accuracy of 98.52% ± 3.00% for origin discrimination. Although practical packaging-level implementation requires further validation of printing quality, substrate readability, scanning robustness, decoder compatibility, and user-facing application design, the generated QR codes demonstrate the potential to provide smartphone-accessible analytical information for future product traceability and quality assessment. This framework provides a compact and reliable proof-of-concept strategy for NMR data storage, food traceability, and consumer-oriented quality control of L. barbarum and other food matrices.