THREE-DIMENSIONAL POINT CLOUD LABEL LEARNING DEVICE, THREE-DIMENSIONAL POINT CLOUD LABEL ESTIMATING DEVICE, METHOD, AND PROGRAM

摘要

A clustering unit (101) divides an inputted three-dimensional point cloud into a plurality of clusters and outputs cluster data. On the basis of the three-dimensional point cloud and the cluster data, a surrounding point sampling unit (102) extracts, for each of the plurality of clusters, a surrounding three-dimensional point cloud that is present within a prescribed distance from the cluster. Accepting as inputs extended cluster data that includes information about the three-dimensional point cloud included in each of the divided clusters and information about the extracted surrounding three-dimensional point cloud, and a correct answer label that indicates an object to which the three-dimensional point cloud included in each of the clusters belongs, a learning unit (103) learns a DNN parameter for estimating the label of each of the clusters from the extended cluster data. An estimation unit (104) inputs extended cluster data pertaining to clusters for which the label is unknown to the DNN in which the parameter was learned and estimates the label of each of the clusters.