Please use this identifier to cite or link to this item:
|Title:||Paired cross-modal data augmentation for fine-grained image-to-text retrieval||Authors:||Wang, Hao
Hoi, Steven C. H.
|Keywords:||Engineering::Computer science and engineering||Issue Date:||2022||Source:||Wang, H., Lin, G., Hoi, S. C. H. & Miao, C. (2022). Paired cross-modal data augmentation for fine-grained image-to-text retrieval. 30th ACM International Conference on Multimedia (MM 2022), 5517-5526. https://dx.doi.org/10.1145/3503161.3547809||Project:||AISG-GC-2019-003
|Abstract:||This paper investigates an open research problem of generating text-image pairs to improve the training of fine-grained image-to-text cross-modal retrieval task, and proposes a novel framework for paired data augmentation by uncovering the hidden semantic information of StyleGAN2 model. Specifically, we first train a StyleGAN2 model on the given dataset. We then project the real images back to the latent space of StyleGAN2 to obtain the latent codes. To make the generated images manipulatable, we further introduce a latent space alignment module to learn the alignment between StyleGAN2 latent codes and the corresponding textual caption features. When we do online paired data augmentation, we first generate augmented text through random token replacement, then pass the augmented text into the latent space alignment module to output the latent codes, which are finally fed to StyleGAN2 to generate the augmented images. We evaluate the efficacy of our augmented data approach on two public cross-modal retrieval datasets, in which the promising experimental results demonstrate the augmented text-image pair data can be trained together with the original data to boost the image-to-text cross-modal retrieval performance.||URI:||https://hdl.handle.net/10356/164145||ISBN:||9781450392037||DOI:||10.1145/3503161.3547809||Rights:||© 2022 The owner/author(s). Publication rights licensed to ACM. All rights reserved.||Fulltext Permission:||none||Fulltext Availability:||No Fulltext|
|Appears in Collections:||SCSE Conference Papers|
Updated on Mar 20, 2023
Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.