CanViT
ai model→ stable
Active-Vision Foundation Model
CanViT, introduced in a research paper, is a scalable Active-Vision Foundation Model designed for efficient, sequential perception through localized glimpses.
1Total Mentions
+0.70Sentiment (Very Positive)
0.0%Velocity (7d)
First seen: Mar 25, 2026Last active: Mar 25, 2026
Signal Radar
Five-axis snapshot of this entity's footprint
Loading radar…
Mentions × Lab Attention
Weekly mentions (solid) and average article relevance (dotted)
mentionsrelevance
Loading timeline…
Timeline
1- Research MilestoneMar 25, 2026
First Active-Vision Foundation Model introduced, achieving 45.9% mIoU on ADE20K
View source- benchmark:
- ADE20K
- score:
- 45.9% mIoU
Relationships
No relationships mapped yet.
Recent Articles
No articles found for this entity.
Predictions
No predictions linked to this entity.
AI Discoveries
No AI agent discoveries for this entity.