US 11,704,841 B2
Apparatus for estimating sameness of point cloud data and system for estimating sameness of point cloud data
Tomohiro Mashita, Osaka (JP); Yuki Uranishi, Osaka (JP); Photchara Ratsamee, Osaka (JP); Kenshiro Tamata, Osaka (JP); Hiromi Ohkubo, Osaka (JP); and Tadafumi Nishimura, Osaka (JP)
Assigned to DAIKIN INDUSTRIES, LTD., Osaka (JP)
Appl. No. 17/754,219
Filed by DAIKIN INDUSTRIES, LTD., Osaka (JP)
PCT Filed Sep. 17, 2020, PCT No. PCT/JP2020/035231
§ 371(c)(1), (2) Date Mar. 28, 2022,
PCT Pub. No. WO2021/065538, PCT Pub. Date Apr. 8, 2021.
Claims priority of application No. JP2019-181004 (JP), filed on Sep. 30, 2019.
Prior Publication US 2022/0343553 A1, Oct. 27, 2022
Int. Cl. G06K 9/00 (2022.01); G06T 9/00 (2006.01); G06N 3/08 (2023.01)
CPC G06T 9/002 (2013.01) [G06N 3/08 (2013.01); G06F 2218/08 (2023.01); G06F 2218/12 (2023.01)] 7 Claims
OG exemplary drawing
 
1. A point cloud data sameness estimation apparatus for estimating sameness of objects that are sources of two 3-dimensional point cloud datasets, comprising:
a memory;
a processor coupled to the memory:
a first neural network configured to output a first point cloud data feature, with information about the first point cloud data as an input into the first neural network; and
a second neural network configured to output a second point cloud data feature, with information about the second point cloud data as an input into the second neural network;
wherein a weight is mutually shared by the first neural network and the second neural network,
wherein the processor is configured to:
acquire the first point cloud data feature, the second point cloud data feature and a label, the first point cloud data feature and the second point cloud data feature each including 3-dimensional point cloud data, the label indicating whether or not the first point cloud data and the second point cloud data are generated respectively based on a same 3-dimensional shape of an object,
output an evaluation about sameness of the first point cloud data and the second point cloud data, based on the first point cloud data feature and the second point cloud data feature,
calculate a gap between the evaluation and the label, and
update the weight based on the gap, and wherein
in a case when the first point cloud data and the second point cloud data are generated based on first 3-dimensional shape and second 3-dimensional shape, respectively, the first 3-dimensional shape being same as the second 3-dimensional shape, information about a point cloud data set of the first point cloud data and the second point cloud data is first training data labeled by the label as being identical, and
in a case when the first point cloud data and the second point cloud data are generated based on first 3-dimensional shape and second 3-dimensional shape, respectively, the first 3-dimensional shape being different from the second 3-dimensional shape, information about the point cloud data set of the first point cloud data and the second point cloud data is second training data labeled by the label as being non-identical.