基于全局学习扩展的可见光-红外行人重识别
A person re-identification method for visible-infrared cross-modal settings, recommended by ChinaMM 2025.
%E9%83%AD%E5%AD%90%E5%BC%BA
A person re-identification method for visible-infrared cross-modal settings, recommended by ChinaMM 2025.
A high-fidelity facial age editing network recommended by CCF CAD/CG 2025.
A text-free, zero-shot speech emotion transfer method based on hierarchical disentanglement.
A hierarchical point cloud analysis and denoising network based on state space models.
A few-shot semantic segmentation method based on mutual-guided conditional coupling.
An attribute-aware dynamic octree network for semantic understanding of point clouds.
A multi-guided foreground enhancement method for person re-identification, recommended by GDC 2024.
A face editing method based on semantic disentanglement and feature fusion, nominated for the ChinaMM 2024 Best Paper/Student Paper award and recommended by the conference.
An RGB-D salient object detection network based on multilevel perception.
A digital elevation model generation framework constrained by feature lines and elevation ranges.