Heterogeneous Domain Adaptation by Information Capturing and Distribution Matching

Heterogeneous domain adaptation (HDA) is a challenging problem because of the different feature representations in the source and target domains. Most HDA methods search for mapping matrices from the source and target domains to discover latent features for learning. However, these methods barely co...

Ausführliche Beschreibung

Bibliographische Detailangaben
Veröffentlicht in:IEEE transactions on image processing : a publication of the IEEE Signal Processing Society. - 1992. - 30(2021) vom: 08., Seite 6364-6376
1. Verfasser: Wu, Hanrui (VerfasserIn)
Weitere Verfasser: Zhu, Hong, Yan, Yuguang, Wu, Jiaju, Zhang, Yifan, Ng, Michael K
Format: Online-Aufsatz
Sprache:English
Veröffentlicht: 2021
Zugriff auf das übergeordnete Werk:IEEE transactions on image processing : a publication of the IEEE Signal Processing Society
Schlagworte:Journal Article
LEADER 01000naa a22002652 4500
001 NLM327829281
003 DE-627
005 20231225201640.0
007 cr uuu---uuuuu
008 231225s2021 xx |||||o 00| ||eng c
024 7 |a 10.1109/TIP.2021.3094137  |2 doi 
028 5 2 |a pubmed24n1092.xml 
035 |a (DE-627)NLM327829281 
035 |a (NLM)34236965 
040 |a DE-627  |b ger  |c DE-627  |e rakwb 
041 |a eng 
100 1 |a Wu, Hanrui  |e verfasserin  |4 aut 
245 1 0 |a Heterogeneous Domain Adaptation by Information Capturing and Distribution Matching 
264 1 |c 2021 
336 |a Text  |b txt  |2 rdacontent 
337 |a ƒaComputermedien  |b c  |2 rdamedia 
338 |a ƒa Online-Ressource  |b cr  |2 rdacarrier 
500 |a Date Revised 14.07.2021 
500 |a published: Print-Electronic 
500 |a Citation Status PubMed-not-MEDLINE 
520 |a Heterogeneous domain adaptation (HDA) is a challenging problem because of the different feature representations in the source and target domains. Most HDA methods search for mapping matrices from the source and target domains to discover latent features for learning. However, these methods barely consider the reconstruction error to measure the information loss during the mapping procedure. In this paper, we propose to jointly capture the information and match the source and target domain distributions in the latent feature space. In the learning model, we propose to minimize the reconstruction loss between the original and reconstructed representations to preserve information during transformation and reduce the Maximum Mean Discrepancy between the source and target domains to align their distributions. The resulting minimization problem involves two projection variables with orthogonal constraints that can be solved by the generalized gradient flow method, which can preserve orthogonal constraints in the computational procedure. We conduct extensive experiments on several image classification datasets to demonstrate that the effectiveness and efficiency of the proposed method are better than those of state-of-the-art HDA methods 
650 4 |a Journal Article 
700 1 |a Zhu, Hong  |e verfasserin  |4 aut 
700 1 |a Yan, Yuguang  |e verfasserin  |4 aut 
700 1 |a Wu, Jiaju  |e verfasserin  |4 aut 
700 1 |a Zhang, Yifan  |e verfasserin  |4 aut 
700 1 |a Ng, Michael K  |e verfasserin  |4 aut 
773 0 8 |i Enthalten in  |t IEEE transactions on image processing : a publication of the IEEE Signal Processing Society  |d 1992  |g 30(2021) vom: 08., Seite 6364-6376  |w (DE-627)NLM09821456X  |x 1941-0042  |7 nnns 
773 1 8 |g volume:30  |g year:2021  |g day:08  |g pages:6364-6376 
856 4 0 |u http://dx.doi.org/10.1109/TIP.2021.3094137  |3 Volltext 
912 |a GBV_USEFLAG_A 
912 |a SYSFLAG_A 
912 |a GBV_NLM 
912 |a GBV_ILN_350 
951 |a AR 
952 |d 30  |j 2021  |b 08  |h 6364-6376