Tu8 - Comprehensive Data Fusion of Nano-FTIR Spectral Orders for Enhanced Identification of Influenza and SARS-CoV-2

In fifteen words or less, explain the significance of this contribution (Novel Aspect).: Systematic analysis of different strategies for combining nano-FTIR phase and absorption across multiple orders.

Abstract Text: Identifying the potential of nano-FTIR for studying viruses offers important insights into the characteristics of viral particles. In the study, we investigated the classification of particles utilizing three orders of phase and absorption spectra incorporating data fusion techniques [1] where absorption derived from the amplitude spectrum. Initially, we analysed orders of phase and absorption spectra individually (single-mode). Then, we studied their combination in two fusion approaches: phase fusion, where we combined all orders of phase spectra, and absorption fusion, where we combined all orders of absorption spectra.The study was conducted in different ways, where we analyzed data at the spectrum level with 3x3 grids per particle and then applied majority voting to obtain particle-level predictions. We also investigated the averaged spectrum as a particle-level approach. We performed preprocessing, where baseline correction and vector normalization were applied to the phase spectra, followed by single-mode and combined analysis. After preprocessing, we combined the data and built partial least square (PLS) model evaluated with particle-wise 5-fold nested cross-validation (CV) [2] where parameter optimization was performed in the inner CV and model testing in the outer CV. Evaluation of combined and single-mode approaches showed that the data fusion outperformed the single-mode orders. However, phase fusion, combining all orders of phase on the particle-level, achieved the best result among all investigated strategies, with a balanced accuracy of 99%.

ACKNOWLEDGMENTS
Financial support from the EU, the TMWWDG, the TAB, the BMBF, the DFG, the Carl-Zeiss Foundation, and the Leibniz Association is greatly acknowledged. This work is supported by the BMBF, funding program Photonics Research Germany (13N15466 (LPI-BT1-FSU), 13N15715 (LPI-BT4-FSU) and 13N15710 (LPI-BT3-FSU)) and is integrated into the Leibniz Center for Photonics in Infection Research (LPI). The LPI initiated by Leibniz-IPHT, Leibniz-HKI, Friedrich Schiller University Jena, and Jena University Hospital is part of the BMBF national roadmap for research infrastructures.

REFERENCES
[1] Azam, Kazi Sultana Farhana, Oleg Ryabchykov, and Thomas Bocklitz. "A Review on Data Fusion of Multidimensional Medical and Biomedical Data." Molecules 27.21 (2022): 7448.
[2] Guo, Shuxia, et al. "Common mistakes in cross-validating classification models." Analytical Methods 9.30 (2017): 4410-4417.