Abstract
Point clouds with rich local geometric information have potentially huge implications in several applications, especially in areas of robotic manipulation and autonomous driving. However, most point cloud processing methods cannot extract enough geometric features from a raw point cloud, which restricts the performance of their downstream tasks such as point cloud classification, shape retrieval and part segmentation. In this paper, the authors propose a new method where a convolution based on geometric primitives is adopted to accurately represent the elusive shape in the form of a point cloud to fully extract hidden geometric features. The key idea of the proposed approach is building a brand-new convolution net named ResSANet on the basis of geometric primitives to learn hierarchical geometry information. Two different modules are devised in our network, Res-SA and ResSA2, to achieve feature fusion at different levels in ResSANet. This work achieves classification accuracy up to 93.2% on the ModelNet40 dataset and the shape retrieval with an effect of 87.4%. The part segmentation experiment also achieves an accuracy of 83.3% (class mIoU) and 85.3% (instance mIoU) on ShapeNet dataset. It is worth mentioning that the number of parameters in this work is just 1.04 M while the network depth is minimal. Experimental results and comparisons with state-of-the-art methods demonstrate that our approach can achieve superior performance.
Highlights
With the rapid development of 3D acquisition technologies, high-precision point clouds are available
Point cloud representation preserves the original geometric information in a 3D space, for example, which enables point clouds that are versatile in many fields, including autonomous driving [1] and robotic manipulation [2]
In [3], the authors propose the pioneering work PointNet, which is able to work on irregular point clouds directly to learn per-point features using shared Multi-Layer Perceptron (MLP) and global features using a symmetrical pooling function
Summary
With the rapid development of 3D acquisition technologies, high-precision point clouds are available. Assumed that geometric information might be implicitly learned directly from the coordinates They proposed Geo-Conv to explicitly model the geometric structure amongst points throughout the hierarchy of feature extraction, which is applied to each point in which the local spherical neighborhood is determined by a radius. It is effective, deficiencies still exist in the above-mentioned methods. Two point-based skip connection modules are devised in the network, Res-SA and Res-SA-2, which can fuse multi-level features to raise accuracy and efficiency in Sensors 2021, 21, 3227 points and KNN points like (b). Two point-based skip connection modules are devised in the network, Res-SA and Res-SA-2, which can fuse multi-level features to raise accuracy and efficiency in point-cloud processing; 3. The results demonstrate they achieve state-of-the-art performance on challenging benchmark datasets, ModelNet40 [10] and ShapeNet [11], across three tasks, i.e., classification, shape retrieval and part segmentation
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
