logo
CanyamAI summaries for academic research
language
MobileMobile
  • Phone FeatureWeChat Mini Program
EnglishEnglish
noticeNotice
BlogBlog
delete
arrow
Return

GraPLUS: Graph-based Placement Using Semantics for image composition

delete2025-06-20
delete0
PRE
AI
M
Mehran Safayani
A
Abdolreza Mirzaei
DOI：10.1016/j.cviu.2025.104427delete
deleteOriginal
deleteOriginal request for help
deleteShare
deleteSave
Abstract

Abstract

En 中文
• Scene graphs and large language models enhance object placement in images. • GPT-2 embeddings capture semantic relationships for better object positioning. • Edge-aware graph neural networks process semantic relationships effectively. • Human evaluators preferred our placement method in 51.8% of test cases. • Superior accuracy (92.1%) with competitive visual quality (FID score 28.83).

Journal

Computer Vision and Image Understanding cover
Computer Vision and Image Understanding
IF:
3.5
Papers:
428
Citations:
7.3K

Organization

No organization information available