Classifiers trained on given databases perform poorly when tested on data acquired in different settings. This is explained in domain adaptation through a shift among distributions of the source and target domains. Attempts to align them have traditionally resulted in works reducing the domain shift by introducing appropriate loss terms, measuring the discrepancies between source and target distributions, in the objective function. Here we take a different route, proposing to align the learned representations by embedding in any given network specific Domain Alignment Layers, designed to match the source and target feature distributions to a reference one. Opposite to previous works which define a priori in which layers adaptation should be performed, our method is able to automatically learn the degree of feature alignment required at different levels of the deep network. Thorough experiments on different public benchmarks, in the unsupervised setting, confirm the power of our approach.

AutoDIAL: Automatic DomaIn Alignment Layers / Carlucci, FABIO MARIA; Porzi, Lorenzo; Caputo, Barbara; Ricci, Elisa; ROTA BULO', Samuel. - ELETTRONICO. - (2017), pp. 5067-5075. (Intervento presentato al convegno 16th IEEE International Conference on Computer Vision (ICCV) tenutosi a Venice, ITALY nel Ottobre 2017) [10.1109/ICCV.2017.542].

AutoDIAL: Automatic DomaIn Alignment Layers

Fabio Maria Carlucci
;
PORZI, LORENZO;Barbara Caputo;Elisa Ricci;ROTA BULO', SAMUEL
2017

Abstract

Classifiers trained on given databases perform poorly when tested on data acquired in different settings. This is explained in domain adaptation through a shift among distributions of the source and target domains. Attempts to align them have traditionally resulted in works reducing the domain shift by introducing appropriate loss terms, measuring the discrepancies between source and target distributions, in the objective function. Here we take a different route, proposing to align the learned representations by embedding in any given network specific Domain Alignment Layers, designed to match the source and target feature distributions to a reference one. Opposite to previous works which define a priori in which layers adaptation should be performed, our method is able to automatically learn the degree of feature alignment required at different levels of the deep network. Thorough experiments on different public benchmarks, in the unsupervised setting, confirm the power of our approach.
2017
16th IEEE International Conference on Computer Vision (ICCV)
computer science; computer vision and pattern recognition
04 Pubblicazione in atti di convegno::04b Atto di convegno in volume
AutoDIAL: Automatic DomaIn Alignment Layers / Carlucci, FABIO MARIA; Porzi, Lorenzo; Caputo, Barbara; Ricci, Elisa; ROTA BULO', Samuel. - ELETTRONICO. - (2017), pp. 5067-5075. (Intervento presentato al convegno 16th IEEE International Conference on Computer Vision (ICCV) tenutosi a Venice, ITALY nel Ottobre 2017) [10.1109/ICCV.2017.542].
File allegati a questo prodotto
File Dimensione Formato  
Carlucci_AutoDIAL_2017.pdf

solo gestori archivio

Tipologia: Versione editoriale (versione pubblicata con il layout dell'editore)
Licenza: Tutti i diritti riservati (All rights reserved)
Dimensione 265.73 kB
Formato Adobe PDF
265.73 kB 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/11573/1016397
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 218
  • ???jsp.display-item.citation.isi??? 184
social impact