Unlike images which are represented in regular dense grids, 3D point clouds are irregular and unordered, hence applying convolution on them can be difficult. In this paper, we extend the dynamic filter to a new convolution operation, named PointConv. PointConv can be applied on point clouds to build deep convolutional networks. We treat convolution kernels as nonlinear functions of the local coordinates of 3D points comprised of weight and density functions. With respect to a given point, the weight functions are learned with multi-layer perceptron networks and density functions through kernel density estimation. The most important contribution of this work is a novel reformulation proposed for efficiently computing the weight functions, which allowed us to dramatically scale up the network and significantly improve its performance. The learned convolution kernel can be used to compute translation-invariant and permutation-invariant convolution on any point set in the 3D space. Besides, PointConv can also be used as deconvolution operators to propagate features from a subsampled point cloud back to its original resolution. Experiments on ModelNet40, ShapeNet, and ScanNet show that deep convolutional neural networks built on PointConv are able to achieve state-of-the-art on challenging semantic segmentation benchmarks on 3D point clouds. Besides, our experiments converting CIFAR-10 into a point cloud showed that networks built on PointConv can match the performance of convolutional networks in 2D images of a similar structure.

No abstract

We propose a novel end-to-end deep scene flow model, called PointPWC-Net, on 3D point clouds in a coarse-tofine fashion. Flow computed at the coarse level is upsampled and warped to a finer level, enabling the algorithm to accommodate for large motion without a prohibitive search space. We introduce novel cost volume, upsampling, and warping layers to efficiently handle 3D point cloud data. Unlike traditional cost volumes that require exhaustively computing all the cost values on a high-dimensional grid, our point-based formulation discretizes the cost volume onto input 3D points, and a PointConv operation efficiently computes convolutions on the cost volume. Experiment results on FlyingThings3D outperform the state-of-the-art by a large margin. We further explore novel self-supervised losses to train our model and achieve comparable results to state-of-the-art trained with supervised loss. Without any fine-tuning, our method also shows great generalization ability on KITTI Scene Flow 2015 dataset, outperforming all previous methods.

SARS-CoV-2, the causative agent for COVID-19, infect human mainly via respiratory tract, which is heavily inhabited by local microbiota. However, the interaction between SARS-CoV-2 and nasopharyngeal microbiota, and the association with metabolome has not been well characterized. Here, metabolomic analysis of blood, urine, and nasopharyngeal swabs from a group of COVID-19 and non-COVID-19 patients, and metagenomic analysis of pharyngeal samples were used to identify the key features of COVID-19. Results showed lactic acid, l -proline, and chlorogenic acid methyl ester (CME) were significantly reduced in the sera of COVID-19 patients compared with non-COVID-19 ones. Nasopharyngeal commensal bacteria including Gemella morbillorum , Gemella haemolysans and Leptotrichia hofstadii were notably depleted in the pharynges of COVID-19 patients, while Prevotella histicola , Streptococcus sanguinis , and Veillonella dispar were relatively increased. The abundance of G. haemolysans and L. hofstadii were significantly positively associated with serum CME, which might be an anti-SARS-CoV-2 bacterial metabolite. This study provides important information to explore the linkage between nasopharyngeal microbiota and disease susceptibility. The findings were based on a very limited number of patients enrolled in this study; a larger size of cohort will be appreciated for further investigation.

No abstract