Abstract
Despite the significant advancements in computer vision models, their ability to generalize to novel object-attribute compositions remains limited. Existing methods for Compositional Zero-Shot Learning (CZSL) mainly focus on image classification. This paper aims to enhance CZSL in object detection without forgetting prior learned knowledge. We use Grounding DINO and incorporate Compositional Soft Prompting (CSP) into it and extend it with Compositional Anticipation. We achieve a 70.5% improvement over CSP on the harmonic mean (HM) between seen and unseen compositions on the CLEVR dataset. Furthermore, we introduce Contrastive Prompt Tuning to incrementally address model confusion between similar compositions. We demonstrate the effectiveness of this method and achieve an increase of 14.5% in HM across the pretrain, increment, and unseen sets. Collectively, these methods provide a framework for learning various compositions with limited data, as well as improving the performance of underperforming compositions when additional data becomes available.
Original language | English |
---|---|
Title of host publication | Pattern Recognition |
Subtitle of host publication | Proceedings of the 27th International Conference, ICPR 2024 |
Editors | Apostolos Antonacopoulos, Subhasis Chaudhuri, Rama Chellappa, Cheng-Lin Liu, Saumik Bhattacharya, Umapada Pal |
Publisher | Springer |
Pages | 265-279 |
Number of pages | 15 |
ISBN (Print) | 978-3-0317-8112-4 |
DOIs | |
Publication status | Published - 2025 |
Event | 27th International Conference on Pattern Recognition, ICPR 2024 - Kolkata, India Duration: 1 Dec 2024 → 5 Dec 2024 |
Publication series
Name | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
---|---|
Volume | 15330 LNCS |
ISSN (Print) | 0302-9743 |
ISSN (Electronic) | 1611-3349 |
Conference
Conference | 27th International Conference on Pattern Recognition, ICPR 2024 |
---|---|
Country/Territory | India |
City | Kolkata |
Period | 1/12/24 → 5/12/24 |
Bibliographical note
Green Open Access added to TU Delft Institutional Repository 'You share, we take care!' - Taverne project https://www.openaccess.nl/en/you-share-we-take-careOtherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.
Keywords
- compositional zero-shot learning
- incremental learning
- prompt tuning