MTSS replaces monolithic video captions with factorized streams and relational grounding, yielding reported gains in understanding benchmarks and generation consistency.
Ingredients: Blending custom photos with video diffusion transformers.arXiv preprint arXiv:2501.01790, 2025a
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
citation-role summary
other 1
citation-polarity summary
fields
cs.CV 3years
2026 3verdicts
UNVERDICTED 3roles
other 1polarities
unclear 1representative citing papers
Proposes IaD framework with Identity Decoupling Loss and Text Alignment Loss for richer, identity-consistent IPT2V without subject-specific fine-tuning.
ARGUS converts MLLM-selected identity evidence into a synchronized 3x3 mosaic injected as negative-time memory in a diffusion model, plus supporting training techniques, to achieve SOTA subject preservation on human video benchmarks.
citing papers explorer
-
Script-a-Video: Deep Structured Audio-visual Captions via Factorized Streams and Relational Grounding
MTSS replaces monolithic video captions with factorized streams and relational grounding, yielding reported gains in understanding benchmarks and generation consistency.
-
Customizing Video Portraits via Identity-ActionDecoupling
Proposes IaD framework with Identity Decoupling Loss and Text Alignment Loss for richer, identity-consistent IPT2V without subject-specific fine-tuning.
-
ARGUS: Stacked Multi-View Identity Mosaic Injection for Subject-Preserving Video Generation
ARGUS converts MLLM-selected identity evidence into a synchronized 3x3 mosaic injected as negative-time memory in a diffusion model, plus supporting training techniques, to achieve SOTA subject preservation on human video benchmarks.