7 papers
PhysFlow: Frequency Decoupled with Dual-Field Rectified Flow for Remote Photoplethysmography
Zixu Li, jianjun Qian, Hang Shao +2
Remote Photoplethysmography (rPPG) enables contactless pulse estimation from facial videos, serving as a vital tool for health monitoring. However, current deep learning methods of…
Self-Consistent Generative Paths via Admissible Random Variational Transport
Lei Luo, Yingzhen Zhang, Jian Yang
Modern generative models often define an entire probability path from a simple prior to the data law, rather than only an endpoint map. Diffusion models follow stochastic denoising…
Conditional Random Ordered Transport Spaces
Lei Luo, Jian Yang
A small Wasserstein distance does not certify that a transformation is admissible. In evidence-constrained, semantic, causal, physical, monotone, or risk-sensitive learning, one mu…
AnchorOPT: Towards Optimizing Dynamic Anchors for Adaptive Prompt Learning
Zheng Li, Yibing Song, Xin Zhang +3
Existing prompt learning methods, which are built upon CLIP models, leverage textual tokens as anchors to guide the learnable soft tokens. This guidance improves CLIP generalizatio…
Controllable-LPMoE: Adapting to Challenging Object Segmentation via Dynamic Local Priors from Mixture-of-Experts
Yanguang Sun, Jiawei Lian, Jian Yang +1
Large-scale foundation models provide powerful feature representations for downstream object segmentation tasks. However, when adapted to specific tasks through the full-parameter…
Diff-Reg v2: Diffusion-Based Matching Matrix Estimation for Image Matching and 3D Registration
Qianliang Wu, Haobo Jiang, Yaqing Ding +5
Establishing reliable correspondences is crucial for all registration tasks, including 2D image registration, 3D point cloud registration, and 2D-3D image-to-point cloud registrati…