Lecture
Data fusion in food analysis
- at -
- ICM Saal 4b
- Type: Lecture
Lecture description
Food analysis is challenged by intrinsic heterogeneity, i.e. variations across geographic origin, harvest, processing, storage and even within individual samples. That degrades reproducibility and complicates together with spurious measurements/outliers and
constraints on storage/transmission to a central intelligence node, reliable automated food quality analysis. To address these issues, robust data-fusion pipelines should combine careful preprocessing and outlier handling, calibration to align instrumentspecific responses, data-level fusion, feature-level fusion exemplified by dynamic feature selection, or higher-level aggregation such as ensemble or decision-level fusion. Minding early- or late fusion architectures, no single fusion method fits all cases.
Effective solutions must be tailored to sensor types, dataset scale and operational constraints to deliver reliable, scalable food assessment.
Literature:
[1] Judith Müller-Maatsch, Francesca Romana Bertani, Arianna Mencattini, Annamaria Gerardino, Eugenio Martinelli, Yannick Weesepoel, Saskia van Ruth, The spectral treasure house of miniaturized instruments for food safety, quality and authenticity applications: A perspective, Trends in Food Science & Technology, Volume 110, 2021, Pages 841-848, https://doi.org/10.1016/j.tifs.2021.01.091.
constraints on storage/transmission to a central intelligence node, reliable automated food quality analysis. To address these issues, robust data-fusion pipelines should combine careful preprocessing and outlier handling, calibration to align instrumentspecific responses, data-level fusion, feature-level fusion exemplified by dynamic feature selection, or higher-level aggregation such as ensemble or decision-level fusion. Minding early- or late fusion architectures, no single fusion method fits all cases.
Effective solutions must be tailored to sensor types, dataset scale and operational constraints to deliver reliable, scalable food assessment.
Literature:
[1] Judith Müller-Maatsch, Francesca Romana Bertani, Arianna Mencattini, Annamaria Gerardino, Eugenio Martinelli, Yannick Weesepoel, Saskia van Ruth, The spectral treasure house of miniaturized instruments for food safety, quality and authenticity applications: A perspective, Trends in Food Science & Technology, Volume 110, 2021, Pages 841-848, https://doi.org/10.1016/j.tifs.2021.01.091.
