In this setting, we show our best testing results. An updated ArXiv preprint is available here. This repository is a reproduced work, and we released a pre-trained network model with 88.0% instance-mean-iou and 86.5% class-mean-iou. Training/Testing setting: We are terribly sorry that we missed our CVPR 2020 code submission.With the help of the Delaunay triangulation for graph construction from point clouds and a multi-scale U-Net for segmentation, we manage to demonstrate the state-of-the-art performance on ShapeNet and PartNet, respectively, with significant improvement over the literature. To accelerate the computation in practice, we further propose a novel hierarchical approximate algorithm. To this end, we are motivated by graph drawing and reformulate it as an integer programming problem to learn the topology-preserving graph-to-grid mapping for each individual point cloud. In contrast to the literature where local patterns in 3D point clouds are captured by customized convolutional operators, in this paper we study the problem of how to effectively and efficiently project such point clouds into a 2D image space so that traditional 2D convolutional neural networks (CNNs) such as U-Net can be applied for segmentation. Learning to Segment 3D Point Clouds in 2D Image Space Overview
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |