Please use this identifier to cite or link to this item:
Title: Performance evaluation of classical classifiers and deep learning approaches for polymers classification based on hyperspectral images
Authors: Lorenzo-Navarro, Javier 
Serranti, Silvia
Bonifazi, Giuseppe
Capobianco, Giuseppe
UNESCO Clasification: 120304 Inteligencia artificial
Keywords: Deep Learning
Hyperspectral Images
Machine Learning
Polymer Classification
Issue Date: 2021
Publisher: Springer 
Project: Evaluación del impacto de microplásticos y contaminantes emergentes en las costas de la Macaronesia 
Journal: Lecture Notes in Computer Science 
Conference: 16th International Work-Conference on Artificial Neural Networks, IWANN 2021 
Abstract: Plastics are very valuable material for their desirable characteristics being one of them, their durability. But this characteristic turns plastics into an environmental problem when they end in the environment, and they become one source of contamination that can last for centuries. Thus, the first step for effective recycling is to identify correctly the types of plastics. In this paper, different classical classifiers as Random Forest, KNN, or SVM are compared with 1-D CNN and LSTM to classify plastics from hyperspectral images. Also, Partial Least Squares Discriminant Analysis has been included as the baseline because is one of the most widely used classifiers in the field of the Chemometrics community. The images were preprocessed with several techniques as Standard Normal Variate or Savitzky-Golay Polynomial Derivative to compare their effectiveness with raw data with the classifiers. The experiments were carried out using hyperspectral images with a 240 bands spectrum, and six types of polymers were considered (PE, PA, PP, PS, PVC, EPS). The best results were obtained with SVM+RBF and 1-D CNN with an accuracy of 99.41% and 99.31% respectively, preprocessing the images previously with Standard Normal Variate. Also, PCA and t-SNE methods were tested for dimensionality reduction, but they don’t improve the classifier performance.
ISBN: 978-3-030-85098-2
ISSN: 0302-9743
DOI: 10.1007/978-3-030-85099-9_23
Source: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) [ISSN 0302-9743], v. 12862 LNCS, p. 281-292, (Enero 2021)
Appears in Collections:Actas de congresos
Show full item record

Google ScholarTM




Export metadata

Items in accedaCRIS are protected by copyright, with all rights reserved, unless otherwise indicated.