PriFit: Learning to Fit Primitives Improves Few Shot Point Cloud Segmentation

Simple item page
dc.contributor.authorSharma, Gopalen_US
dc.contributor.authorDash, Bidyaen_US
dc.contributor.authorRoyChowdhury, Arunien_US
dc.contributor.authorGadelha, Matheusen_US
dc.contributor.authorLoizou, Mariosen_US
dc.contributor.authorCao, Liangliangen_US
dc.contributor.authorWang, Ruien_US
dc.contributor.authorLearned-Miller, Erik G.en_US
dc.contributor.authorMaji, Subhransuen_US
dc.contributor.authorKalogerakis, Evangelosen_US
dc.contributor.editorCampen, Marcelen_US
dc.contributor.editorSpagnuolo, Michelaen_US
dc.date.accessioned2022-06-27T16:19:49Z
dc.date.available2022-06-27T16:19:49Z
dc.date.issued2022
dc.description.abstractWe present PRIFIT, a semi-supervised approach for label-efficient learning of 3D point cloud segmentation networks. PRIFIT combines geometric primitive fitting with point-based representation learning. Its key idea is to learn point representations whose clustering reveals shape regions that can be approximated well by basic geometric primitives, such as cuboids and ellipsoids. The learned point representations can then be re-used in existing network architectures for 3D point cloud segmentation, and improves their performance in the few-shot setting. According to our experiments on the widely used ShapeNet and PartNet benchmarks, PRIFIT outperforms several state-of-the-art methods in this setting, suggesting that decomposability into primitives is a useful prior for learning representations predictive of semantic parts. We present a number of ablative experiments varying the choice of geometric primitives and downstream tasks to demonstrate the effectiveness of the method.en_US
dc.description.number5
dc.description.sectionheadersLearning and Creating
dc.description.seriesinformationComputer Graphics Forum
dc.description.volume41
dc.identifier.doi10.1111/cgf.14601
dc.identifier.issn1467-8659
dc.identifier.pages39-50
dc.identifier.pages12 pages
dc.identifier.urihttps://doi.org/10.1111/cgf.14601
dc.identifier.urihttps://diglib.eg.org:443/handle/10.1111/cgf14601
dc.publisherThe Eurographics Association and John Wiley & Sons Ltd.en_US
dc.subjectCCS Concepts: Computing methodologies --> Shape representations; Neural networks; Theory of computation --> Semi-supervised learning
dc.subjectComputing methodologies
dc.subjectShape representations
dc.subjectNeural networks
dc.subjectTheory of computation
dc.subjectSemi
dc.subjectsupervised learning
dc.titlePriFit: Learning to Fit Primitives Improves Few Shot Point Cloud Segmentationen_US
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
v41i5pp039-050.pdf
Size:
3.45 MB
Format:
Adobe Portable Document Format
Download
Collections
41-Issue 5