Enhancing the performance of point cloud completion assisted with tactile information

Point cloud completion is a crucial task in the field of 3D imaging and is becoming a current research hotspot. However, traditional point cloud completion methods often fail to accurately restore the local details of objects due to the lack of detailed descriptions of the missing parts. This paper proposes a performance enhancement scheme for point cloud completion using tactile information. By employing a U-Net-based tactile feature extraction and fusion module, we enhance point cloud completion through the integration of tactile information. We validated our proposed network on the 3DVT dataset, and the experimental results demonstrate that our method effectively improves the performance of point cloud completion. Additionally, we conducted simulation experiments to mimic real-world scenarios where a robotic arm contacts objects, obtaining tactile point clouds using a mounted tactile simulator. These tactile point clouds were then used in related point cloud completion experiments, and the results were evaluated. The experimental results indicate that our proposed method significantly enhances the performance of point cloud completion and improves the local detail of the completed point clouds.