Colors characterize each object around us. For this reason, the study of colors has played a key role in Artificial Intelligence (think, for instance, of image classification, object recognition and segmentation). However, there are some topics about colors still little explored. One of them concerns fabric colors. This is a particular topic since fabrics have some characteristics, such as specific textures, that are not found in other contexts. In this paper, we want to propose a new Convolutional Neural Network (CNN) based model for identifying fabric colors. After introducing this model, we consider three different versions of it and create an ensemble of the corresponding CNNs to get better results. Finally, through a series of experiments, we show that our ensemble is able to improve the state-of-the-art on the identification of fabric colors.

Defining a deep neural network ensemble for identifying fabric colors

Amelio A.
Primo
;
2022-01-01

Abstract

Colors characterize each object around us. For this reason, the study of colors has played a key role in Artificial Intelligence (think, for instance, of image classification, object recognition and segmentation). However, there are some topics about colors still little explored. One of them concerns fabric colors. This is a particular topic since fabrics have some characteristics, such as specific textures, that are not found in other contexts. In this paper, we want to propose a new Convolutional Neural Network (CNN) based model for identifying fabric colors. After introducing this model, we consider three different versions of it and create an ensemble of the corresponding CNNs to get better results. Finally, through a series of experiments, we show that our ensemble is able to improve the state-of-the-art on the identification of fabric colors.
File in questo prodotto:
File Dimensione Formato  
ColorRecognition (3).pdf

accesso aperto

Tipologia: Documento in Pre-print
Dimensione 5.48 MB
Formato Adobe PDF
5.48 MB Adobe PDF Visualizza/Apri
1-s2.0-S1568494622007360-main.pdf

Solo gestori archivio

Tipologia: PDF editoriale
Dimensione 3.6 MB
Formato Adobe PDF
3.6 MB Adobe PDF   Visualizza/Apri   Richiedi una copia

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11564/799717
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 16
  • ???jsp.display-item.citation.isi??? 15
social impact