LoGo: Local-Global Rewards for Consistent Long-Horizon Video Generation okładka

LoGo: Local-Global Rewards for Consistent Long-Horizon Video Generation

LoGo: Local-Global Rewards for Consistent Long-Horizon Video Generation

Słuchaj bezpłatnie, bez subskrypcji

Zobacz szczegóły

Saisonale Sparangebote | 0,99 € pro Monat für die ersten 3 Monate

Danach 6,99 €/Monat. Bedingungen gelten. Monatlich kündbar.
Camera-controlled video models struggle with 3D inconsistency over long horizons: objects vanish or scene structure shifts as the camera moves. Post-training with a single scalar reward gives poor credit assignment for these local errors. LoGo blends global rewards, which preserve camera following and video quality, with spatially localized rewards that give fine-grained feedback. Tested on three base models, DL3DV, and the new TrajectoryBench for long, complex camera paths, it reduces object shifts, artifacts, and scene changes. Authors: Ziqi Ma, Shreya Sharma, Mohamed El Banani, Katja Schwarz, Chongjie Ye, Chao-Yuan Wu, Li Fei-Fei, Ben Mildenhall, Georgia Gkioxari, Justin Johnson, Gowthami Somepalli Paper: https://arxiv.org/abs/2610.03636v1
adbl_web_anon_alc_button_suppression_t1
Brak recenzji