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


Current browse context:


Change to browse by:


References & Citations

DBLP - CS Bibliography


(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: Modeling Image Composition for Complex Scene Generation

Abstract: We present a method that achieves state-of-the-art results on challenging (few-shot) layout-to-image generation tasks by accurately modeling textures, structures and relationships contained in a complex scene. After compressing RGB images into patch tokens, we propose the Transformer with Focal Attention (TwFA) for exploring dependencies of object-to-object, object-to-patch and patch-to-patch. Compared to existing CNN-based and Transformer-based generation models that entangled modeling on pixel-level&patch-level and object-level&patch-level respectively, the proposed focal attention predicts the current patch token by only focusing on its highly-related tokens that specified by the spatial layout, thereby achieving disambiguation during training. Furthermore, the proposed TwFA largely increases the data efficiency during training, therefore we propose the first few-shot complex scene generation strategy based on the well-trained TwFA. Comprehensive experiments show the superiority of our method, which significantly increases both quantitative metrics and qualitative visual realism with respect to state-of-the-art CNN-based and transformer-based methods. Code is available at this https URL
Comments: CVPR 2022
Subjects: Computer Vision and Pattern Recognition (cs.CV)
Cite as: arXiv:2206.00923 [cs.CV]
  (or arXiv:2206.00923v1 [cs.CV] for this version)

Submission history

From: Zuopeng Yang [view email]
[v1] Thu, 2 Jun 2022 08:34:25 GMT (13412kb,D)

Link back to: arXiv, form interface, contact.