We present a simple and general framework for feature learning from point
cloud. The key to the success of CNNs is the convolution operator that is
capable of leveraging spatially-local correlation in data represented densely
in grids (e.g. images). However, point cloud are irregular and unordered, thus
a direct convolving of kernels against the features associated with the points
will result in deserting the shape information while being variant to the
orders. To address these problems, we propose to learn a X-transformation from
the input points, and then use it to simultaneously weight the input features
associated with the points and permute them into latent potentially canonical
order, before the element-wise product and sum operations are applied. The
proposed method is a generalization of typical CNNs into learning features from
point cloud, thus we call it PointCNN. Experiments show that PointCNN achieves
on par or better performance than state-of-the-art methods on multiple
challenging benchmark datasets and tasks.
%0 Generic
%1 li2018pointcnn
%A Li, Yangyan
%A Bu, Rui
%A Sun, Mingchao
%A Chen, Baoquan
%D 2018
%K CNN point_cloud
%T PointCNN
%U http://arxiv.org/abs/1801.07791
%X We present a simple and general framework for feature learning from point
cloud. The key to the success of CNNs is the convolution operator that is
capable of leveraging spatially-local correlation in data represented densely
in grids (e.g. images). However, point cloud are irregular and unordered, thus
a direct convolving of kernels against the features associated with the points
will result in deserting the shape information while being variant to the
orders. To address these problems, we propose to learn a X-transformation from
the input points, and then use it to simultaneously weight the input features
associated with the points and permute them into latent potentially canonical
order, before the element-wise product and sum operations are applied. The
proposed method is a generalization of typical CNNs into learning features from
point cloud, thus we call it PointCNN. Experiments show that PointCNN achieves
on par or better performance than state-of-the-art methods on multiple
challenging benchmark datasets and tasks.
@misc{li2018pointcnn,
abstract = {We present a simple and general framework for feature learning from point
cloud. The key to the success of CNNs is the convolution operator that is
capable of leveraging spatially-local correlation in data represented densely
in grids (e.g. images). However, point cloud are irregular and unordered, thus
a direct convolving of kernels against the features associated with the points
will result in deserting the shape information while being variant to the
orders. To address these problems, we propose to learn a X-transformation from
the input points, and then use it to simultaneously weight the input features
associated with the points and permute them into latent potentially canonical
order, before the element-wise product and sum operations are applied. The
proposed method is a generalization of typical CNNs into learning features from
point cloud, thus we call it PointCNN. Experiments show that PointCNN achieves
on par or better performance than state-of-the-art methods on multiple
challenging benchmark datasets and tasks.},
added-at = {2018-02-10T13:43:14.000+0100},
author = {Li, Yangyan and Bu, Rui and Sun, Mingchao and Chen, Baoquan},
biburl = {https://www.bibsonomy.org/bibtex/2a18e59b707a098dcb0bfc12894c3e137/jk_itwm},
description = {[1801.07791] PointCNN},
interhash = {52048af3e658b0e985c0fa13a688f87e},
intrahash = {a18e59b707a098dcb0bfc12894c3e137},
keywords = {CNN point_cloud},
note = {cite arxiv:1801.07791Comment: Small updates},
timestamp = {2018-02-10T13:43:14.000+0100},
title = {PointCNN},
url = {http://arxiv.org/abs/1801.07791},
year = 2018
}