A CVAE-based Variational Information Flow module is proposed to counteract visual attenuation in MLLMs and improve fine-grained perception on VQA and grounding tasks.
InEuro- pean conference on computer vision, pages 235–251
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
From Attenuation to Attention: Variational Information Flow Manipulation for Fine-Grained Visual Perception
A CVAE-based Variational Information Flow module is proposed to counteract visual attenuation in MLLMs and improve fine-grained perception on VQA and grounding tasks.