We propose a new representation of distance information that is independent from any specific acquisition device, based on the size of portrayed subjects. In this alternative description, each pixel of an image is associated with the size, in real life, of what it represents. Using our proposed representation, datasets acquired with different devices can be effortlessly combined to build more powerful models, and monocular distance estimation can be performed on images acquired from devices that were never used during training. To assess the advantages of the proposed representation, we used it to train a fully convolutional neural network that predicts with pixel-precision the size of different subjects depicted in the image, as a proxy for their distance. Experimental results show that our representation, allowing the combination of heterogeneous training datasets, makes it possible for the trained network to gain better results at test time.

Bianco, S., Buzzelli, M., Schettini, R. (2019). A unifying representation for pixel-precise distance estimation. MULTIMEDIA TOOLS AND APPLICATIONS, 78(10), 13767-13786 [10.1007/s11042-018-6568-2].

A unifying representation for pixel-precise distance estimation

Bianco, Simone;Buzzelli, Marco
;
Schettini, Raimondo
2019

Abstract

We propose a new representation of distance information that is independent from any specific acquisition device, based on the size of portrayed subjects. In this alternative description, each pixel of an image is associated with the size, in real life, of what it represents. Using our proposed representation, datasets acquired with different devices can be effortlessly combined to build more powerful models, and monocular distance estimation can be performed on images acquired from devices that were never used during training. To assess the advantages of the proposed representation, we used it to train a fully convolutional neural network that predicts with pixel-precision the size of different subjects depicted in the image, as a proxy for their distance. Experimental results show that our representation, allowing the combination of heterogeneous training datasets, makes it possible for the trained network to gain better results at test time.
Articolo in rivista - Articolo scientifico
Convolutional neural network; Depth estimation; Distance estimation; Perspective geometry;
Distance estimation; Depth estimation; Perspective geometry; Convolutional neural network
English
24-ago-2018
2019
78
10
13767
13786
reserved
Bianco, S., Buzzelli, M., Schettini, R. (2019). A unifying representation for pixel-precise distance estimation. MULTIMEDIA TOOLS AND APPLICATIONS, 78(10), 13767-13786 [10.1007/s11042-018-6568-2].
File in questo prodotto:
File Dimensione Formato  
2018d_CAMERA_A_unifying_representation_for_pixel_precise_distance_estimation.pdf

Solo gestori archivio

Descrizione: Articolo principale
Tipologia di allegato: Publisher’s Version (Version of Record, VoR)
Dimensione 4.48 MB
Formato Adobe PDF
4.48 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/10281/205398
Citazioni
  • Scopus 7
  • ???jsp.display-item.citation.isi??? 5
Social impact