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

Abstract
We 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.
Description

CCS Concepts: Computing methodologies --> Shape representations; Neural networks; Theory of computation --> Semi-supervised learning

        
@article{
10.1111:cgf.14601
, journal = {Computer Graphics Forum}, title = {{
PriFit: Learning to Fit Primitives Improves Few Shot Point Cloud Segmentation
}}, author = {
Sharma, Gopal
and
Dash, Bidya
and
RoyChowdhury, Aruni
and
Gadelha, Matheus
and
Loizou, Marios
and
Cao, Liangliang
and
Wang, Rui
and
Learned-Miller, Erik G.
and
Maji, Subhransu
and
Kalogerakis, Evangelos
}, year = {
2022
}, publisher = {
The Eurographics Association and John Wiley & Sons Ltd.
}, ISSN = {
1467-8659
}, DOI = {
10.1111/cgf.14601
} }
Citation
Collections