Rethinking Generalized Zero-Shot Learning : A Synthesized Per-Instance Attribute Perspective

Generalized zero-shot learning (GZSL) shows great potential for improving generalization to unseen classes in real-world scenarios. However, most GZSL methods depend on benchmark datasets with per-class attribute annotations, which creates a large semantic gap and worsens the domain shift problem in...

Description complète

Détails bibliographiques
Publié dans:IEEE transactions on image processing : a publication of the IEEE Signal Processing Society. - 1992. - 34(2025) vom: 12., Seite 5847-5859
Auteur principal: Tang, Chenwei (Auteur)
Autres auteurs: Wang, Ying, Xie, Wei, Zhang, Qianjun, Xiao, Rong, He, Zhenan, Lv, Jiancheng
Format: Article en ligne
Langue:English
Publié: 2025
Accès à la collection:IEEE transactions on image processing : a publication of the IEEE Signal Processing Society
Sujets:Journal Article
LEADER 01000caa a22002652c 4500
001 NLM392639793
003 DE-627
005 20250923233015.0
007 cr uuu---uuuuu
008 250917s2025 xx |||||o 00| ||eng c
024 7 |a 10.1109/TIP.2025.3607612  |2 doi 
028 5 2 |a pubmed25n1578.xml 
035 |a (DE-627)NLM392639793 
035 |a (NLM)40953421 
040 |a DE-627  |b ger  |c DE-627  |e rakwb 
041 |a eng 
100 1 |a Tang, Chenwei  |e verfasserin  |4 aut 
245 1 0 |a Rethinking Generalized Zero-Shot Learning  |b A Synthesized Per-Instance Attribute Perspective 
264 1 |c 2025 
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 22.09.2025 
500 |a published: Print 
500 |a Citation Status PubMed-not-MEDLINE 
520 |a Generalized zero-shot learning (GZSL) shows great potential for improving generalization to unseen classes in real-world scenarios. However, most GZSL methods depend on benchmark datasets with per-class attribute annotations, which creates a large semantic gap and worsens the domain shift problem in the visual-semantic space. To address these challenges, instance-level attributes offer an intuitive solution, but they require expensive manual annotation. In this paper, we propose a simple yet effective approach called per-instance attribute synthesis (PIAS) to generate diverse semantic representations for each instance. Our method first uses the Vision Transformer (ViT) model to extract visual features and then generates per-instance attributes. The patch splitting, positional embedding, and multi-head self-attention mechanisms in ViT improve the discriminability of both visual and semantic representations. Next, we define the generated attributes of class-average images as class anchor points. These anchor points are calibrated in the semantic space by minimizing the cosine similarity between the anchor points and per-class attribute annotations. Finally, we improve the diversity of generated per-instance attributes by aligning the topological structure between per-class attribute annotations and synthesized per-instance attributes with that between class-average visual features and per-instance visual features. We conduct comprehensive experiments on three challenging ZSL datasets: AWA2, CUB, and SUN. The results show that PIAS significantly outperforms state-of-the-art methods under both ZSL and GZSL settings. We further demonstrate the generalization ability of PIAS by applying it to attribute-based zero-shot image retrieval tasks 
650 4 |a Journal Article 
700 1 |a Wang, Ying  |e verfasserin  |4 aut 
700 1 |a Xie, Wei  |e verfasserin  |4 aut 
700 1 |a Zhang, Qianjun  |e verfasserin  |4 aut 
700 1 |a Xiao, Rong  |e verfasserin  |4 aut 
700 1 |a He, Zhenan  |e verfasserin  |4 aut 
700 1 |a Lv, Jiancheng  |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 34(2025) vom: 12., Seite 5847-5859  |w (DE-627)NLM09821456X  |x 1941-0042  |7 nnas 
773 1 8 |g volume:34  |g year:2025  |g day:12  |g pages:5847-5859 
856 4 0 |u http://dx.doi.org/10.1109/TIP.2025.3607612  |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 34  |j 2025  |b 12  |h 5847-5859