Unsupervised domain adaptation for object detection addresses the adaption of detectors trained in a source domain to work accurately in an unseen target domain. Recently, methods approaching the alignment of the intermediate features proven to be promising, achieving state-of-the-art results. However, these methods are laborious to implement and hard to interpret. Although promising, there is still room for improvements to close the performance gap toward the upper-bound (when training with the target data). In this work, we propose a method to generate an artificial dataset in the target domain to train an object detector. We employed two unsupervised image translators (CycleGAN and an AdaIN-based model) using only annotated data from the source domain and non-annotated data from the target domain. Our key contributions are the proposal of a less complex yet more effective method that also has an improved interpretability. Results on real-world scenarios for autonomous driving show significant improvements, outperforming state-of-the-art methods in most cases, further closing the gap toward the upper-bound.

Cross-domain object detection using unsupervised image translation / Arruda, V. F.; Berriel, R. F.; Paixao, T. M.; Badue, C.; De Souza, A. F.; Sebe, N.; Oliveira-Santos, T.. - In: EXPERT SYSTEMS WITH APPLICATIONS. - ISSN 0957-4174. - 192:(2022), pp. 11633401-11633410. [10.1016/j.eswa.2021.116334]

Cross-domain object detection using unsupervised image translation

Sebe N.;
2022-01-01

Abstract

Unsupervised domain adaptation for object detection addresses the adaption of detectors trained in a source domain to work accurately in an unseen target domain. Recently, methods approaching the alignment of the intermediate features proven to be promising, achieving state-of-the-art results. However, these methods are laborious to implement and hard to interpret. Although promising, there is still room for improvements to close the performance gap toward the upper-bound (when training with the target data). In this work, we propose a method to generate an artificial dataset in the target domain to train an object detector. We employed two unsupervised image translators (CycleGAN and an AdaIN-based model) using only annotated data from the source domain and non-annotated data from the target domain. Our key contributions are the proposal of a less complex yet more effective method that also has an improved interpretability. Results on real-world scenarios for autonomous driving show significant improvements, outperforming state-of-the-art methods in most cases, further closing the gap toward the upper-bound.
2022
Arruda, V. F.; Berriel, R. F.; Paixao, T. M.; Badue, C.; De Souza, A. F.; Sebe, N.; Oliveira-Santos, T.
Cross-domain object detection using unsupervised image translation / Arruda, V. F.; Berriel, R. F.; Paixao, T. M.; Badue, C.; De Souza, A. F.; Sebe, N.; Oliveira-Santos, T.. - In: EXPERT SYSTEMS WITH APPLICATIONS. - ISSN 0957-4174. - 192:(2022), pp. 11633401-11633410. [10.1016/j.eswa.2021.116334]
File in questo prodotto:
File Dimensione Formato  
Arruda-ESA22.pdf

Solo gestori archivio

Tipologia: Versione editoriale (Publisher’s layout)
Licenza: Tutti i diritti riservati (All rights reserved)
Dimensione 2.12 MB
Formato Adobe PDF
2.12 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/11572/340560
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 6
  • ???jsp.display-item.citation.isi??? 5
social impact