Convolutional Neural Networks (CNNs) trained on large scale RGB databases have become the secret sauce in the majority of recent approaches for object categorization from RGB-D data. Thanks to colorization techniques, these methods exploit the filters learned from 2D images to extract meaningful representations in 2.5D. Still, the perceptual signature of these two kind of images is very different, with the first usually strongly characterized by textures, and the second mostly by silhouettes of objects. Ideally, one would like to have two CNNs, one for RGB and one for depth, each trained on a suitable data collection, able to capture the perceptual properties of each channel for the task at hand. This has not been possible so far, due to the lack of a suitable depth database. This paper addresses this issue, proposing to opt for synthetically generated images rather than collecting by hand a 2.5D large scale database. While being clearly a proxy for real data, synthetic images allow to trade quality for quantity, making it possible to generate a virtually infinite amount of data. We show that the filters learned from such data collection, using the very same architecture typically used on visual data, learns very different filters, resulting in depth features (a) able to better characterize the different facets of depth images, and (b) complementary with respect to those derived from CNNs pre-trained on 2D datasets. Experiments on two publicly available databases show the power of our approach.

A deep representation for depth images from synthetic data / Carlucci, FABIO MARIA; Russo, Paolo; Caputo, Barbara. - ELETTRONICO. - (2017), pp. 1362-1369. (Intervento presentato al convegno 2017 IEEE International Conference on Robotics and Automation, ICRA 2017 tenutosi a Singapore; Singapore) [10.1109/ICRA.2017.7989162].

A deep representation for depth images from synthetic data

CARLUCCI, FABIO MARIA
;
RUSSO, PAOLO;CAPUTO, BARBARA
2017

Abstract

Convolutional Neural Networks (CNNs) trained on large scale RGB databases have become the secret sauce in the majority of recent approaches for object categorization from RGB-D data. Thanks to colorization techniques, these methods exploit the filters learned from 2D images to extract meaningful representations in 2.5D. Still, the perceptual signature of these two kind of images is very different, with the first usually strongly characterized by textures, and the second mostly by silhouettes of objects. Ideally, one would like to have two CNNs, one for RGB and one for depth, each trained on a suitable data collection, able to capture the perceptual properties of each channel for the task at hand. This has not been possible so far, due to the lack of a suitable depth database. This paper addresses this issue, proposing to opt for synthetically generated images rather than collecting by hand a 2.5D large scale database. While being clearly a proxy for real data, synthetic images allow to trade quality for quantity, making it possible to generate a virtually infinite amount of data. We show that the filters learned from such data collection, using the very same architecture typically used on visual data, learns very different filters, resulting in depth features (a) able to better characterize the different facets of depth images, and (b) complementary with respect to those derived from CNNs pre-trained on 2D datasets. Experiments on two publicly available databases show the power of our approach.
2017
2017 IEEE International Conference on Robotics and Automation, ICRA 2017
Computer Science; Computer Vision and Pattern Recognition
04 Pubblicazione in atti di convegno::04b Atto di convegno in volume
A deep representation for depth images from synthetic data / Carlucci, FABIO MARIA; Russo, Paolo; Caputo, Barbara. - ELETTRONICO. - (2017), pp. 1362-1369. (Intervento presentato al convegno 2017 IEEE International Conference on Robotics and Automation, ICRA 2017 tenutosi a Singapore; Singapore) [10.1109/ICRA.2017.7989162].
File allegati a questo prodotto
File Dimensione Formato  
Carlucci_Preprint_A-deep_2017.pdf

accesso aperto

Note: https://ieeexplore.ieee.org/document/7989162
Tipologia: Documento in Pre-print (manoscritto inviato all'editore, precedente alla peer review)
Licenza: Tutti i diritti riservati (All rights reserved)
Dimensione 2.53 MB
Formato Adobe PDF
2.53 MB Adobe PDF
Carlucci_A-deep_2017.pdf

solo gestori archivio

Tipologia: Versione editoriale (versione pubblicata con il layout dell'editore)
Licenza: Tutti i diritti riservati (All rights reserved)
Dimensione 1.53 MB
Formato Adobe PDF
1.53 MB Adobe PDF   Contatta l'autore

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/11573/965602
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 23
  • ???jsp.display-item.citation.isi??? ND
social impact