The proposed Coordinate-Aware Feature Excitation (CAFE) module and Position-Aware Upsampling (Pos-Up) module both adhere to ...
Abstract: We introduce Point-LN, a novel lightweight frame-work engineered for efficient 3D point cloud classification. Point-LN integrates essential non-parametric components-such as Farthest Point ...
Abstract: Vision-language foundation models (VLMs) have shown great potential in feature transfer and generalization across a wide spectrum of medical-related downstream tasks. However, fine-tuning ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results