A position paper arguing that vision foundation models need dynamic relational graphs for relational reasoning, with evidence drawn from the author's own prior action recognition and tumor segmentation systems.
Laptev, On space-time interest points, in: Proceedings of the IEEE International Conference on Computer Vision (ICCV), IEEE, 2005, pp
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Why Relational Graphs Will Save the Next Generation of Vision Foundation Models?
A position paper arguing that vision foundation models need dynamic relational graphs for relational reasoning, with evidence drawn from the author's own prior action recognition and tumor segmentation systems.