Null-Space Tuning injects learnable residuals into input features confined to the null-space for high-quality inputs to preserve pre-trained knowledge while directing restoration components for low-quality inputs outside that space.
Object-aware multi-branch relation networks for spatio-temporal video grounding
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
fields
cs.CV 3years
2026 3verdicts
UNVERDICTED 3representative citing papers
Bridge-STG decouples spatio-temporal alignment via semantic bridging and query-guided localization modules to achieve state-of-the-art m_vIoU of 34.3 on VidSTG among MLLM methods.
A new large-scale subjective database for video portrait region cropping with temporal smoothing, benchmarked using existing models and compared to saliency predictions.
citing papers explorer
-
Knowledge-Preserved Model Tuning in Null-Space for Robust Spatio-Temporal Video Grounding
Null-Space Tuning injects learnable residuals into input features confined to the null-space for high-quality inputs to preserve pre-trained knowledge while directing restoration components for low-quality inputs outside that space.
-
Bridging Time and Space: Decoupled Spatio-Temporal Alignment for Video Grounding
Bridge-STG decouples spatio-temporal alignment via semantic bridging and query-guided localization modules to achieve state-of-the-art m_vIoU of 34.3 on VidSTG among MLLM methods.
-
Subjective Portrait Region Cropping in Landscape Videos with Temporal Annotation Smoothing
A new large-scale subjective database for video portrait region cropping with temporal smoothing, benchmarked using existing models and compared to saliency predictions.