Semi-supervised 3D object recognition through CNN labeling

Please use this identifier to cite or link to this item: http://hdl.handle.net/10045/74443
Información del item - Informació de l'item - Item information
Title: Semi-supervised 3D object recognition through CNN labeling
Authors: Rangel, José Carlos | Martínez-Gómez, Jesús | Romero-González, Cristina | García-Varea, Ismael | Cazorla, Miguel
Research Group/s: Robótica y Visión Tridimensional (RoViT)
Center, Department or Service: Universidad de Alicante. Departamento de Ciencia de la Computación e Inteligencia Artificial | Universidad de Alicante. Instituto Universitario de Investigación Informática
Keywords: Object recognition | Deep learning | Object labeling | Machine learning
Knowledge Area: Ciencia de la Computación e Inteligencia Artificial
Issue Date: Apr-2018
Publisher: Elsevier
Citation: Applied Soft Computing. 2018, 65: 603-613. doi:10.1016/j.asoc.2018.02.005
Abstract: Despite the outstanding results of Convolutional Neural Networks (CNNs) in object recognition and classification, there are still some open problems to address when applying these solutions to real-world problems. Specifically, CNNs struggle to generalize under challenging scenarios, like recognizing the variability and heterogeneity of the instances of elements belonging to the same category. Some of these difficulties are directly related to the input information, 2D-based methods still show a lack of robustness against strong lighting variations, for example. In this paper, we propose to merge techniques using both 2D and 3D information to overcome these problems. Specifically, we take advantage of the spatial information in the 3D data to segment objects in the image and build an object classifier, and the classification capabilities of CNNs to semi-supervisedly label each object image for training. As the experimental results demonstrate, our model can successfully generalize for categories with high intra-class variability and outperform the accuracy of a well-known CNN model.
Sponsor: This work has been partially sponsored by the Spanish Ministry of Economy and Competitiveness under Grant Number TIN2015-65686-C5-3-R. It has been also supported by the Spanish Government TIN2016-76515-R Grant, supported with Feder funds. Cristina Romero-González is funded by the MECD Grant FPU12/04387. José Carlos Rangel is funded by the IFARHU Grant 8-2014-166 of the Republic of Panamá.
URI: http://hdl.handle.net/10045/74443
ISSN: 1568-4946 (Print) | 1872-9681 (Online)
DOI: 10.1016/j.asoc.2018.02.005
Language: eng
Type: info:eu-repo/semantics/article
Rights: © 2018 Elsevier B.V.
Peer Review: si
Publisher version: https://doi.org/10.1016/j.asoc.2018.02.005
Appears in Collections:INV - RoViT - Artículos de Revistas

Files in This Item:
Files in This Item:
File Description SizeFormat 
Thumbnail2018_Rangel_etal_ApplSoftComp_final.pdfVersión final (acceso restringido)3,91 MBAdobe PDFOpen    Request a copy
Thumbnail2018_Rangel_etal_ApplSoftComp_accepted.pdfEmbargo 24 meses (acceso abierto: 8 febr. 2020)5,16 MBAdobe PDFOpen    Request a copy


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