2017/04/04 by Roman Klokov, Klokov, Roman, Victor Lempitsky +1 · 1 citation
Computer Science · Earth and Planetary Sciences · Engineering · #3D Shape Modeling and Analysis #3D Surveying and Cultural Heritage #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Image Processing and 3D Reconstruction
paper · pdf · doi:10.48550/arxiv.1704.01222
openalex publication_date 2017/04/04 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
We present a new deep learning architecture (called Kd-network) that is\ndesigned for 3D model recognition tasks and works with unstructured point\nclouds. The new architecture performs multiplicative transformations and share\nparameters of these transformations according to the subdivisions of the point\nclouds imposed onto them by Kd-trees. Unlike the currently dominant\nconvolutional architectures that usually require rasterization on uniform\ntwo-dimensional or three-dimensional grids, Kd-networks do not rely on such\ngrids in any way and therefore avoid poor scaling behaviour. In a series of\nexperiments with popular shape recognition benchmarks, Kd-networks demonstrate\ncompetitive performance in a number of shape recognition tasks such as shape\nclassification, shape retrieval and shape part segmentation.\n