By decomposing slot channels into static and dynamic subspaces via CCD and using CTR for cross-temporal feature reconstruction, the method internalizes temporal consistency without explicit SSC loss.
Smoothing Slot Attention Iterations and Recurrences
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
abstract
Slot Attention (SA) lies at the heart of mainstream Object-Centric Learning (OCL). Image features can be aggregated into object-level representations by SA \textit{iteratively} refining cold-start query slots. For video, such aggregation proceeds by SA \textit{recurrently} shared across frames, with queries cold-started on the first frame while transitioned from the previous frame's slots thereafter. However, cold-start queries lack sample-specific cues thus hindering precise aggregation on image or video's first frame; Non-first frames' queries are already sample-specific thus requiring aggregation transforms different from the first frame. We address these issues with our \textit{SmoothSA}: (1) To smooth SA iterations on image or video's first frame, we \textit{preheat} cold-start queries with rich input-feature information, by a tiny module self-distilled inside OCL; (2) To smooth SA recurrences across video's first and non-first frames, we \textit{differentiate} the homogeneous aggregation transforms by using full and single iterations respectively. Comprehensive experiments on object discovery, recognition and visual reasoning validate our method's effectiveness. Further visual analyses illuminate the underline mechanisms. Our \textit{source code}, \textit{model checkpoints} and \textit{training logs} are provided on https://github.com/Genera1Z/SmoothSA.
fields
cs.CV 2years
2026 2verdicts
UNVERDICTED 2representative citing papers
Implicit cycle consistency on the reconstruction manifold avoids feature collapse in video object-centric learning unlike explicit slot-space consistency.
citing papers explorer
-
Internalizing Temporal Consistency in Video Object-Centric Learning without Explicit Regularization
By decomposing slot channels into static and dynamic subspaces via CCD and using CTR for cross-temporal feature reconstruction, the method internalizes temporal consistency without explicit SSC loss.
-
Cycle Consistency in Video Object-Centric Learning
Implicit cycle consistency on the reconstruction manifold avoids feature collapse in video object-centric learning unlike explicit slot-space consistency.