Open access
Nov 2026
Unlocking the Potential of Vision Foundation Models for Multi-Task Dense Prediction via Mixture of Task Adapters
D2MoE is proposed, a Mixture of Task Adapters that explicitly factorizes these roles within a frozen DINOv3 backbone, establishing a principled and scalable route to prior-preserving, foundation-centric multi-task dense perception.
Yun-Zhi Zhuge, Xin-Zhuo Yu, Si-Tong Gong et al.
· IEEE Transactions on Pattern... · 0 citations