Get Started
Home
Topics
Search
Library
Research questionHow can image editing shift a source image’s mixed emotion distribution while preserving its scene composition and semantics?A single emotion label cannot specify which coexisting emotions should increase, decrease, or remain stable. Editing affective visual cues can also disrupt the source scene’s content, composition, or structure.
Computer Vision
Diffusion Models
Image & Video Processing
Image Generation
Machine Learning
Latest papersRecent research connected to this question, newest first.AffectDelta: Beyond Emotion Labels for Image EditingThe source presents AffectDelta, a source-aware diffusion editor that represents transitions between eight-dimensional emotion distributions. It uses predicted source states and signed source-to-target differences, and is evaluated on 248,841 predicted-distribution source-target pairs covering cross-category and within-category transitions through quantitative comparisons, qualitative examples, and ablations.research paper · Sep 2, 2026
Related questions
How can image editors infer edit regions and preserve non-target content without introducing artifacts?How can generative image systems preserve subject identity under viewpoint changes, degradation, and iterative edits?How can composed image retrieval distinguish changed, preserved, and removed visual details?How can text-guided image editing explain edits while keeping provenance resistant to white-box removal?