A long standing problem in visual object categorization is the ability of algorithms to generalize across different testing conditions. The problem has been formalized as a covariate shift among the probability distributions generating the training data (source) and the test data (target) and several domain adaptation methods have been proposed to address this issue. While these approaches have considered the single sourcesingle target scenario, it is plausible to have multiple sources and require adaptation to any possible target domain. This last scenario, named Domain Generalization (DG), is the focus of our work. Differently from previous DG methods which learn domain invariant representations from source data, we design a deep network with multiple domain-specific classifiers, each associated to a source domain. At test time we estimate the probabilities that a target sample belongs to each source domain and exploit them to optimally fuse the classifiers predictions. To further improve the generalization ability of our model, we also introduced a domain agnostic component supporting the final classifier. Experiments on two public benchmarks demonstrate the power of our approach.

Best Sources Forward: Domain Generalization through Source-Specific Nets / Mancini, Massimiliano; Bulo, Samuel Rota; Caputo, Barbara; Ricci, Elisa. - (2018), pp. 1353-1357. (Intervento presentato al convegno ICIP tenutosi a Athens, Grece nel 7-10 October 2018) [10.1109/ICIP.2018.8451318].

Best Sources Forward: Domain Generalization through Source-Specific Nets

Mancini, Massimiliano;Ricci, Elisa
2018-01-01

Abstract

A long standing problem in visual object categorization is the ability of algorithms to generalize across different testing conditions. The problem has been formalized as a covariate shift among the probability distributions generating the training data (source) and the test data (target) and several domain adaptation methods have been proposed to address this issue. While these approaches have considered the single sourcesingle target scenario, it is plausible to have multiple sources and require adaptation to any possible target domain. This last scenario, named Domain Generalization (DG), is the focus of our work. Differently from previous DG methods which learn domain invariant representations from source data, we design a deep network with multiple domain-specific classifiers, each associated to a source domain. At test time we estimate the probabilities that a target sample belongs to each source domain and exploit them to optimally fuse the classifiers predictions. To further improve the generalization ability of our model, we also introduced a domain agnostic component supporting the final classifier. Experiments on two public benchmarks demonstrate the power of our approach.
2018
25th IEEE International Conference on Image Processing (ICIP)
Piscataway, NJ USA
IEEE
978-1-4799-7061-2
Mancini, Massimiliano; Bulo, Samuel Rota; Caputo, Barbara; Ricci, Elisa
Best Sources Forward: Domain Generalization through Source-Specific Nets / Mancini, Massimiliano; Bulo, Samuel Rota; Caputo, Barbara; Ricci, Elisa. - (2018), pp. 1353-1357. (Intervento presentato al convegno ICIP tenutosi a Athens, Grece nel 7-10 October 2018) [10.1109/ICIP.2018.8451318].
File in questo prodotto:
File Dimensione Formato  
best.pdf

Solo gestori archivio

Tipologia: Versione editoriale (Publisher’s layout)
Licenza: Tutti i diritti riservati (All rights reserved)
Dimensione 2.16 MB
Formato Adobe PDF
2.16 MB Adobe PDF   Visualizza/Apri

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/11572/225598
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 76
  • ???jsp.display-item.citation.isi??? 52
social impact