NAVE	Networked Augmented Virtual Environment (NAVE) Group

Publication:Weimin Shi, Yuan Xiong, Qianwen Wang, Han Jiang, Zhong Zhou. FDCPNet: Feature Discrimination and Context Propagation Network for 3D Shape Representation[J]. Virtual Reality & Intelligent Hardware (VRIH), 2024. pdf         3D shape representation using mesh data is essential in various applications, such as virtual reality and simulation technologies. Current methods extracting features from mesh edges or faces struggle with complex 3D models due to edge-based approaches missing global context and face-based methods overlooking variations in adjacent areas, which affects overall precision. To address these issues, we propose the Feature Discrimination and Context Propagation Network (FDCPNet), a novel approach that synergistically integrates local and global features in mesh datasets. FDCPNet is composed of two modules: 1) Feature Discrimination Module (FDM), which employs an attention mechanism to enhance the identification of key local features, 2) Context Propagation Module (CPM) enriches key local features by integrating global contextual information, facilitating a more detailed and comprehensive representation of crucial areas within the mesh model. Experiments on the popular datasets validate FDCPNet's effectiveness, showing a improvement in classification accuracy over the baseline MeshNet. Furthermore, even with reduced mesh face numbers and limited training data, FDCPNet achieves promising results, showing its robustness in variable complexity scenarios. create by admin at 2024-07-16 18:31:21