Plenary 2

Spectral Unmixing Revisited: latest trends in multimodal image analysis

Abstract

Spectral unmixing remains a cornerstone in hyperspectral image analysis, enabling the extraction of spectral fingerprints and concentration maps from the image constituents. Although unmixing of individual images is a common practice, there are still many challenges associated with the unmixing task on multimodal image measurements. Spatial and spectral diversity among imaging platforms, such as differences in spatial resolution or area scanned, or in spectral dimensionality, with some imaging techniques that provide a linear spectrum and others a 2D spectroscopic landscape per pixel, do not fit in the classical bilinear unmixing model linked to image measurements.

Adapting to these multimodal challenges passes through exploiting new data configurations that may accommodate all the available original information, even if there is not always an equivalence among all image platforms, and using more flexible models that may incorporate bilinear and higher order models in the same analysis. These new trends in multimodal unmixing are shown with real examples of chemical and biological analysis.

Authors

A. De Juan1 , A. Sicre-Conesa1, M. Marsal2, P. Loza-ALvarez2, R. Rocha de Oliveira1 and A. Gómez-Sánchez1.

Chemometrics group. Universitat de Barcelona.
ICFO, Institute of Photonic Sciences, Barcelona.

References

de Juan, A., de Oliveira, R. R., & Gómez-Sánchez, A. (2024). Multiset analysis by multivariate curve resolution: The unmixing methodology to handle hyperspectral image fusion scenarios. In Data Handling in Science and Technology (Vol. 33, pp. 111-132). Elsevier.
Gómez-Sánchez, A., Marro, M., Marsal, M., Loza-Alvarez, P., & de Juan, A. (2020). 3D and 4D image fusion: Coping with differences in spectroscopic modes among hyperspectral images. Analytical Chemistry, 92(14), 9591-9602.
Sicre-Conesa, A., Marsal, M., Gómez-Sánchez, A., Loza-Álvarez, P., & de Juan, A. (2025). Multiscale and Multimodal Image Fusion. Coping with Differences in Scanned Area and Spatial Resolution for Raman/Fluorescence Images of Labeled Cells. Analytical Chemistry.