We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.CV

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Computer Vision and Pattern Recognition

Title: Word-Level Fine-Grained Story Visualization

Abstract: Story visualization aims to generate a sequence of images to narrate each sentence in a multi-sentence story with a global consistency across dynamic scenes and characters. Current works still struggle with output images' quality and consistency, and rely on additional semantic information or auxiliary captioning networks. To address these challenges, we first introduce a new sentence representation, which incorporates word information from all story sentences to mitigate the inconsistency problem. Then, we propose a new discriminator with fusion features and further extend the spatial attention to improve image quality and story consistency. Extensive experiments on different datasets and human evaluation demonstrate the superior performance of our approach, compared to state-of-the-art methods, neither using segmentation masks nor auxiliary captioning networks.
Comments: ECCV 2022
Subjects: Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL); Machine Learning (cs.LG)
Cite as: arXiv:2208.02341 [cs.CV]
  (or arXiv:2208.02341v3 [cs.CV] for this version)

Submission history

From: Bowen Li [view email]
[v1] Wed, 3 Aug 2022 21:01:47 GMT (5468kb,D)
[v2] Thu, 25 Aug 2022 07:34:24 GMT (5469kb,D)
[v3] Thu, 22 Sep 2022 08:11:44 GMT (5467kb,D)

Link back to: arXiv, form interface, contact.