Synthesize then align: Modality alignment augmentation for zero-shot image captioning with synthetic data
Authors:
- Zhiyue LiuZLZhiyue Liu
- Jinyuan LiuJLJinyuan Liu
- Xin LingXLXin Ling
- Qingbao HuangQHQingbao Huang
- Jiahai WangJWJiahai Wang
- 6 March 2025
- journal article
- Published by Elsevier in Knowledge-Based Systems
- Vol. 315
Abstract
No abstract availableFunding Information
- National Natural Science Foundation of China (62406081)
- National Natural Science Foundation of China (62276072)
- Natural Science Foundation of Guangxi Province (2025GXNSFDA069017)
- Natural Science Foundation of Guangxi Province (2025GXNSFBA069232)
This publication has 8 references indexed in Scilit:
- Diverse and tailored image generation for zero-shot multi-label classificationKnowledge-Based Systems, 2024
- FireDM: A weakly-supervised approach for massive generation of multi-scale and multi-scene fire segmentation datasetsKnowledge-Based Systems, 2024
- Generating More Pertinent Captions by Leveraging Semantics and Style on Multi-Source DatasetsInternational Journal of Computer Vision, 2023
- Multimodal Data Augmentation for Image Captioning using Diffusion Models2023
- Parallel encoder–decoder framework for image captioningKnowledge-Based Systems, 2023
- Prompt-Based Learning for Unpaired Image CaptioningIEEE Transactions on Multimedia, 2023
- Unpaired Image Captioning by Image-Level Weakly-Supervised Visual Concept RecognitionIEEE Transactions on Multimedia, 2022
- From image descriptions to visual denotations: New similarity metrics for semantic inference over event descriptionsTransactions of the Association for Computational Linguistics, 2014