RE-VLM is the first dual-stream VLM combining RGB and event data with a graph-based pipeline to generate training captions and QA pairs, showing gains over RGB-only and event-only models on new datasets for challenging scenes.
Multimodal alzheimer’s disease recognition from image, text and audio.Scientific Reports, 15(1):29038,
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
-
RE-VLM: Event-Augmented Vision-Language Model for Scene Understanding
RE-VLM is the first dual-stream VLM combining RGB and event data with a graph-based pipeline to generate training captions and QA pairs, showing gains over RGB-only and event-only models on new datasets for challenging scenes.