S-MSRRS5000: A Simulated Dataset Highlighting the Challenges of Data Obtained From Multiple Spatially Resolved Reflection Spectroscopy
Date
2024-04-29Author
Magnussen, Birk Martin
Contributing Person/Institution
Researcher: Jessulat, Maik
Supervisor: Stern, Claudius
Supervisor: Sick, Bernhard
Metadata
Show full item recordAbstract
Optical sensors based on spectroscopy are occasionally used in consumer healthcare and wellness applications. This includes applications such as measuring the concentration of cutaneous carotenoids using sensors based on multiple spatially resolved reflection spectroscopy (MSRRS). Processing the data yielded from MSRRS-based sensors poses unique challenges. When using machine learning for data processing, specialized models such as continuous feature networks are required to achieve good results. However, due to privacy issues of medical data, data availability is low, hindering model development.
S-MSRRS5000 is a simulated dataset, highlighting the challenges of data from MSRRS-based sensors. These challenges are presented as four different challenge sets with increasing prediction difficulty, with 5000 simulated measurements per challenge set.
Funder
Hessisches Ministerium für Digitalisierung und InnovationRelated Resources
IsSupplementTo: (URL) https://www.thinkmind.org/index.php?view=article&articleid=intsys_v16_n34_2023_3IsSupplementTo: (DOI) https://doi.org/10.17170/kobra-2024042910096
Collections
Except where otherwise noted, this item's license is described as Creative Commons Attribution-NonCommercial-NoDerivatives 4.0