EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Comparison of Artificial Neural Network and Polynomial Approximation Models for Reflectance Spectra Reconstruction

Mihael Lazar, Aleš Hladnik iD

Sensors · 15.01.2023 · 23 (2) · 1000

DOI: 10.3390/s23021000 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, online erschienen am 15.01.2023. Verfasst von Mihael Lazar und 1 weiteren Autor:innen. Beteiligte Einrichtung: Department of Textiles, Graphic Arts and Design, Faculty of Natural Sciences and Engineering, University of Ljubljana, Snežniška Ulica 5, SI-1000 Lju…. Hauptthemen laut MEDLINE: Algorithms, Neural Networks, Computer. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 33 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Slovenian Research Agency.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Knowledge of surface reflection of an object is essential in many technological fields, including graphics and cultural heritage. Compared to direct multi- or hyper-spectral capturing approaches, commercial RGB cameras allow for a high resolution and fast acquisition, so the idea of mapping this information into a reflectance spectrum (RS) is promising. This study compared two modelling approaches based on a training set of RGB-reflectance pairs, one implementing artificial neural networks (ANN) and the other one using multivariate polynomial approximation (PA). The effect of various parameters was investigated: the ANN learning algorithm—standard backpropagation (BP) or Levenberg-Marquardt (LM), the number of hidden layers (HLs) and neurons, the degree of multivariate polynomials in PA, the number of inputs, and the training set size on both models. In the two-layer ANN with significantly fewer inputs than outputs, a better MSE performance was found where the number of neurons in the first HL was smaller than in the second one. For ANNs with one and two HLs with the same number of neurons in the first layer, the RS reconstruction performance depends on the choice of BP or LM learning algorithm. RS reconstruction methods based on ANN and PA are comparable, but the ANN models’ better fine-tuning capabilities enable, under realistic constraints, finding ANNs that outperform PA models. A profiling approach was proposed to determine the initial number of neurons in HLs—the search centre of ANN models for different training set sizes.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

AlgorithmsNeural Networks, ComputerModels, Statistical

Stichwörter der Autor:innen

artificial neural networkpolynomial approximationreflectance reconstruction

Autor:innen und Institutionen

  • Department of Textiles, Graphic Arts and Design, Faculty of Natural Sciences and Engineering, University of Ljubljana, Snežniška Ulica 5, SI-1000 Ljubljana, Slovenia.

ORCID-Profile: Aleš Hladnik

Förderung

  • Slovenian Research Agency

Zitieren

APA

Lazar, M., & Hladnik, A. (2023). Comparison of Artificial Neural Network and Polynomial Approximation Models for Reflectance Spectra Reconstruction. Sensors, 23(2), 1000. https://doi.org/10.3390/s23021000

Vancouver

Lazar M, Hladnik A. Comparison of Artificial Neural Network and Polynomial Approximation Models for Reflectance Spectra Reconstruction. Sensors. 2023;23(2):1000. doi:10.3390/s23021000

BibTeX
@article{Lazar2023,
  author = {Lazar, Mihael and Hladnik, Aleš},
  title = {Comparison of Artificial Neural Network and Polynomial Approximation Models for Reflectance Spectra Reconstruction},
  journal = {Sensors},
  year = {2023},
  volume = {23},
  number = {2},
  pages = {1000},
  doi = {10.3390/s23021000}
}
Kostenlos starten →