arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2003.13834cs.CVcs.GRcs.LG
收藏

基于近似凸分解的点云标记有效学习

Label-Efficient Learning on Point Clouds using Approximate Convex Decompositions

University of Massachusetts Amherst(马萨诸塞大学阿默斯特分校)

Matheus Gadelha, Aruni RoyChowdhury, Gopal Sharma, Evangelos Kalogerakis, Liangliang Cao, Erik Learned-Miller, Rui Wang, Subhransu Maji

摘要:

从三维点云数据中进行形状分类和零件分割的问题在过去的几年中得到了越来越多的关注。然而,这两个问题都受到相对较小的训练集的影响,需要统计有效的方法来学习3D形状表示。在本文中,我们研究了使用近似凸分解(ACD)作为一个自监督信号的标签有效的学习点云表示。我们表明,使用ACD近似地面实况分割提供了良好的自我监督学习3D点云表示,是非常有效的下游任务。我们报告了ModelNet40形状分类数据集上的无监督表示学习的最新技术水平的改进,以及ShapeNetPart数据集上Few-Shot部分分割的显著增益。代码可在https://github.com/matheusgadelha/PointCloudLearningACD上获得

英文摘要:

The problems of shape classification and part segmentation from 3D point clouds have garnered increasing attention in the last few years. Both of these problems, however, suffer from relatively small training sets, creating the need for statistically efficient methods to learn 3D shape representations. In this paper, we investigate the use of Approximate Convex Decompositions (ACD) as a self-supervisory signal for label-efficient learning of point cloud representations. We show that using ACD to approximate ground truth segmentation provides excellent self-supervision for learning 3D point cloud representations that are highly effective on downstream tasks. We report improvements over the state-of-the-art for unsupervised representation learning on the ModelNet40 shape classification dataset and significant gains in few-shot part segmentation on the ShapeNetPart dataset.Code available at https://github.com/matheusgadelha/PointCloudLearningACD

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑