Composite Convolution: a Flexible Operator for Deep Learning on 3D Point Clouds
This work addresses the challenge of handling scattered and irregular 3D point data for researchers and practitioners in computer vision and robotics, offering a novel operator that improves flexibility and regularization in network design.
The paper tackles the problem of processing 3D point clouds with deep neural networks by introducing a composite layer as a flexible alternative to existing convolutional operators, achieving state-of-the-art performance in anomaly detection and competitive results in classification and segmentation compared to methods like ConvPoint and KPConv.
Deep neural networks require specific layers to process point clouds, as the scattered and irregular location of 3D points prevents the use of conventional convolutional filters. We introduce the composite layer, a flexible and general alternative to the existing convolutional operators that process 3D point clouds. We design our composite layer to extract and compress the spatial information from the 3D coordinates of points and then combine this with the feature vectors. Compared to mainstream point-convolutional layers such as ConvPoint and KPConv, our composite layer guarantees greater flexibility in network design and provides an additional form of regularization. To demonstrate the generality of our composite layers, we define both a convolutional composite layer and an aggregate version that combines spatial information and features in a nonlinear manner, and we use these layers to implement CompositeNets. Our experiments on synthetic and real-world datasets show that, in both classification, segmentation, and anomaly detection, our CompositeNets outperform ConvPoint, which uses the same sequential architecture, and achieve similar results as KPConv, which has a deeper, residual architecture. Moreover, our CompositeNets achieve state-of-the-art performance in anomaly detection on point clouds. Our code is publicly available at \url{https://github.com/sirolf-otrebla/CompositeNet}.