PatchCNN: An Explicit Convolution Operator for Point Clouds Perception

Fei Wang,Xing Zhang,Yong Jiang,Li Kong,Xiaotong Wei

DOI: https://doi.org/10.1109/lgrs.2020.2981507

IF: 5.343

2021-04-01

IEEE Geoscience and Remote Sensing Letters

Abstract:A novel convolution architecture PatchCNN is proposed for extending 2-D grid convolution to the nongrid structured data: point clouds, without any intermediate data representation. Previous studies implicitly capture local shape pattern from the meaningful subset or a local region without considering the interaction among points of the local region. The PointPatch module in our deep network, in spirit to the 8-pixels neighborhood in the 2-D image, explicitly models geometric relationship among points in the local region. We adopt a light 3-D convolution network to adaptively integrate features of the PointPatch module. The integrated features encode geometric relationship and the impact of surrounding points, which brings sufficient shape awareness and robustness for point cloud perception. Additionally, in our work, the convolution weight on each point is treated as a Lipschitz continuous function approximated by multilayer perceptron (MLP) and integrated features in the local region. Theoretically, the explicit learning strategy proposed in PatchCNN introduces inductive bias beneficial to the learning shape pattern in 3-D Euclidean space. Extensive experiments on ModelNet40 and ScanNet v2 data sets demonstrate that the proposed method achieves the competitive performance on par or even better than state-of-the-art methods.

imaging science & photographic technology,remote sensing,engineering, electrical & electronic,geochemistry & geophysics

What problem does this paper attempt to address?

This paper aims to solve the signal integrity problems between vias and traces in multi - layer printed circuit boards (PCBs). Specifically, the paper proposes a hybrid method for analyzing the signal integrity of vias and traces in an infinitely large board - to - board pair. This method separates the via domain and the trace domain from the board domain through the Domain Decomposition Method (DDM). Each independent via and trace structure domain is modeled as a multi - mode network consisting of TEM - mode waveguide ports and vertical lumped ports on the split interface. The board domain is mainly dominated by the TM \(Z_0\) mode, and its network parameters are solved by the Boundary Integral Method (BIM). By connecting the networks of each domain and enforcing the field continuity condition along the split interface, the final S - parameters of the vias and traces and their coupling are obtained. This method can effectively optimize the via structure layout and trace routing, especially when the via mode and trace routing are determined, the coupling problem can be modeled efficiently.