analiza aiqbench
FreeMatching addresses a critical bottleneck in generative workflows by removing rigid spatio-temporal constraints, enabling dense correspondence matching even when physical continuity is broken during image editing or reference-guided generation. By leveraging heterogeneous supervision and teacher-guided refinement, the framework establishes identity-preserving mappings that outperform classical benchmarks while serving as a quantitative metric for human-aligned identity preservation. This advancement directly enhances the fidelity of complex generative tasks where traditional motion assumptions fail.
Beyond Spatio-Temporal Priors: A Generalizable Approach for Dense Correspondence Matching
arXiv:2610.12421v1 Announce Type: cross Abstract: Dense correspondence matching has historically been bounded by simplifying spatio-temporal priors, such as smooth motion and rigid geometry. While effective for classical tasks, these assumptions break down in image editing and reference-guided generation (IEG), where transformations can preserve visual identity while breaking physical continuity. To establish identity-preserving correspondence across such transformations, we introduce FreeMatching, a generalizable framework combining generative and semantic foundation representations with heterogeneous supervision from classical datasets, tracked videos, and synthetic scenes. Teacher-guided iterative refinement further improves correspondence in IEG without dense correspondence annotations. Experimentally, a single FreeMatching model substantially improves correspondence quality on challenging IEG image pairs while retaining competitive performance on classical benchmarks. Furthermore, we demonstrate its utility as a quantitative metric for evaluating identity preservation, with scores that correlate with human judgment. The code is available at https://github.com/luping-liu/FreeMatching.
Abstract:Dense correspondence matching has historically been bounded by simplifying spatio-temporal priors, such as smooth motion and rigid geometry. While effective for classical tasks, these assumptions break down in image editing and reference-guided generation (IEG), where transformations can preserve visual identity while breaking physical continuity. To establish identity-preserving correspondence across such transformations, we introduce FreeMatching, a generalizable framework combining generative and semantic foundation representations with heterogeneous supervision from classical datasets, tracked videos, and synthetic scenes. Teacher-guided iterative refinement further improves correspondence in IEG without dense correspondence annotations. Experimentally, a single FreeMatching model substantially improves correspondence quality on challenging IEG image pairs while retaining competitive performance on classical benchmarks. Furthermore, we demonstrate its utility as a quantitative metric for evaluating identity preservation, with scores that correlate with human judgment. The code is available at this https URL.
Accepted at NeurIPS 2026. 24 pages, 7 figures, including appendices
Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
arXiv:2610.12421 [cs.CV]
(or arXiv:2610.12421v1 [cs.CV] for this version)
https://doi.org/10.48550/arXiv.2610.12421 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 245 cuvinte
Subiecte conexe
Reflections on unlearning and inoculation
LessWrong · 20 sept. 2026
Modernizing and scaling support operations with generative AI on AWS
AWS · 2 sept. 2026
Anticipatory Risk-Guided Reinforcement Learning for Safe Flight Through Dynamic Clutter
arXiv · 28 iul. 2026
SUM: Unified Geometric Surgery on Spatio-Temporal Adaptation Vectors for Federated Class Incremental Learning
arXiv · 23 iul. 2026