Show simple item record

dc.contributor.authorHermosilla Casajús, Pedro
dc.contributor.authorRistchel, Tobias
dc.contributor.authorVázquez Alcocer, Pere Pau
dc.contributor.authorVinacua Pla, Álvaro
dc.contributor.authorRopinski, Timo
dc.contributor.otherUniversitat Politècnica de Catalunya. Departament de Ciències de la Computació
dc.identifier.citationHermosilla, P. [et al.]. Monte Carlo convolution for learning on non-uniformly sampled point clouds. "ACM transactions on graphics", Novembre 2018, vol. 37, núm. 6, p. 235:1-235:12.
dc.description.abstractDeep learning systems extensively use convolution operations to process input data. Though convolution is clearly defined for structured data such as 2D images or 3D volumes, this is not true for other data types such as sparse point clouds. Previous techniques have developed approximations to convolutions for restricted conditions. Unfortunately, their applicability is limited and cannot be used for general point clouds. We propose an efficient and effective method to learn convolutions for non-uniformly sampled point clouds, as they are obtained with modern acquisition techniques. Learning is enabled by four key novelties: first, representing the convolution kernel itself as a multilayer perceptron; second, phrasing convolution as a Monte Carlo integration problem, third, using this notion to combine information from multiple samplings at different levels; and fourth using Poisson disk sampling as a scalable means of hierarchical point cloud learning. The key idea across all these contributions is to guarantee adequate consideration of the underlying non-uniform sample distribution function from a Monte Carlo perspective. To make the proposed concepts applicable to real-world tasks, we furthermore propose an efficient implementation which significantly reduces the GPU memory required during the training process. By employing our method in hierarchical network architectures we can outperform most of the state-of-the-art networks on established point cloud segmentation, classification and normal estimation benchmarks. Furthermore, in contrast to most existing approaches, we also demonstrate the robustness of our method with respect to sampling variations, even when training with uniformly sampled data only. To support the direct application of these concepts, we provide a ready-to-use TensorFlow implementation of these layers at
dc.subjectÀrees temàtiques de la UPC::Informàtica::Intel·ligència artificial::Aprenentatge automàtic
dc.subject.lcshNeural networks (Computer science)
dc.subject.lcshMonte Carlo method
dc.subject.lcshMachine learning
dc.subject.otherDeep learning
dc.subject.otherConvolutional neural networks
dc.subject.otherPoint clouds
dc.subject.otherMonte Carlo integration
dc.titleMonte Carlo convolution for learning on non-uniformly sampled point clouds
dc.subject.lemacXarxes neuronals (Informàtica)
dc.subject.lemacMontecarlo, Mètode de
dc.subject.lemacAprenentatge automàtic
dc.contributor.groupUniversitat Politècnica de Catalunya. ViRVIG - Grup de Recerca en Visualització, Realitat Virtual i Interacció Gràfica
dc.description.peerreviewedPeer Reviewed
dc.rights.accessOpen Access
dc.description.versionPostprint (published version)
upcommons.citation.authorHermosilla, P.; Ristchel, T.; Vázquez, P.; Vinacua, A.; Ropinski, T.
upcommons.citation.publicationNameACM transactions on graphics

Files in this item


This item appears in the following Collection(s)

Show simple item record

All rights reserved. This work is protected by the corresponding intellectual and industrial property rights. Without prejudice to any existing legal exemptions, reproduction, distribution, public communication or transformation of this work are prohibited without permission of the copyright holder