Title:
GCN-BASED THREE-DIMENSIONAL OBJECT RECOGNITION AND POSE ESTIMATION METHOD AND DEVICE
Document Type and Number:
WIPO Patent Application WO/2024/076023
Kind Code:
A1
Abstract:
The present invention relates to a three-dimensional object recognition and pose estimation method performed by means of a three-dimensional object recognition and pose estimation device, the method comprising the steps of: receiving a point cloud; configuring a graph from the point cloud; repeatedly applying a graph convolutional network (GCN) to the graph so as to sense an object and estimate a pose; and determining a bounding box of the sensed object.
Inventors:
JUNG TAE WON (KR)
JUNG KYE DONG (KR)
JUNG CHI SEO (KR)
KIM IN SEON (KR)
YU MIN SU (KR)
PARK CHAN SOO (KR)
KANG JIN KYU (KR)
JUNG KYE DONG (KR)
JUNG CHI SEO (KR)
KIM IN SEON (KR)
YU MIN SU (KR)
PARK CHAN SOO (KR)
KANG JIN KYU (KR)
Application Number:
PCT/KR2023/013667
Publication Date:
April 11, 2024
Filing Date:
September 12, 2023
Export Citation:
Assignee:
TSP XR INC (KR)
International Classes:
G06V20/64; G06N3/08; G06V10/46; G06V10/75; G06V10/762; G06V10/774
Foreign References:
KR20220126934A | 2022-09-19 | |||
KR20220012658A | 2022-02-04 | |||
KR20220038996A | 2022-03-29 | |||
KR20200052420A | 2020-05-15 |
Other References:
정치서. GCN 기반의 3D 포인트 클라우드 객체 검출. June 2021 (JEONG, Chi-Seo. 3D point cloud object detection based on graph convolutional network). [Retrieved on 07 December 2023]. Retrieved from .
Attorney, Agent or Firm:
TNI IP LAW FIRM (KR)
Download PDF: