WO2023277975 SYSTEMS AND METHODS OF CONTRASTIVE POINT COMPLETION WITH FINE-TO-COARSE REFINEMENT

Settings

Stemming

Single Family Member

Include NPL

Feedback

Goto Application

Publication Number WO/2023/277975

Publication Date 05.01.2023

International Application No. PCT/US2022/018281

International Filing Date 01.03.2022

IPC

CPC

View more classifications

View less classifications

Applicants

TENCENT AMERICA LLC [US]/[US]

Inventors

WANG, Bo
CHEN, Weikai
YANG, Bo
LIU, Songrun

Agents

SUN, Yalei
BREGMAN, Dion, M.
CRISMAN, Douglas, J.
CHOU, Catherine, S.
BERNSTEIN, David, P.

Priority Data

17/364,586

30.06.2021

Publication Language English (en)

Filing Language English (EN)

Designated States

View all

Title

(EN) SYSTEMS AND METHODS OF CONTRASTIVE POINT COMPLETION WITH FINE-TO-COARSE REFINEMENT

(FR) SYSTÈMES ET PROCÉDÉS DE COMPLÉTION CONTRASTIVE DE POINTS AVEC RAFFINEMENT FIN-GROSSIER

Abstract

(EN) An electronic apparatus performs a method of generating a complete point cloud of a 3D object from a partial point cloud of the 3D object obtained via raw scanning. The method includes: constructing a second point cloud of the 3D object from the first point cloud using a teacher- student neural network that has a dual network structure comprising a teacher network and a student network both sharing same architecture; and refining the second point cloud by applying symmetry detection and enrichment to the second point cloud to obtain a symmetrical second point cloud of the 3D object and updating the symmetrical second point cloud to get a third point cloud of the 3D object. The teacher-student neural network has a dual network structure comprising a teacher network and a student network both sharing the same architecture.

(FR) Selon l'invention, un appareil électronique réalise un procédé de génération d'un nuage de points complet d'un objet 3D à partir d'un nuage de points partiel de l'objet 3D obtenu via numérisation brute. Le procédé comprend les étapes consistant à: construire un deuxième nuage de points de l'objet 3D à partir du premier nuage de points à l'aide d'un réseau neuronal enseignant-étudiant qui présente une structure à réseaux jumelés comportant un réseau enseignant et un réseau étudiant partageant tous deux une même architecture; et raffiner le deuxième nuage de points en appliquant une détection de symétrie et un enrichissement au deuxième nuage de points pour obtenir un deuxième nuage de points symétrique de l'objet 3D et mettre à jour le deuxième nuage de points symétrique pour obtenir un troisième nuage de points de l'objet 3D. Le réseau neuronal enseignant-étudiant présente une structure à réseaux jumelés comportant un réseau enseignant et un réseau étudiant partageant tous deux la même architecture.

G	PHYSICS
06	COMPUTING; CALCULATING OR COUNTING
K	RECOGNITION OF DATA; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
9	Methods or arrangements for reading or recognising printed or written characters or for recognising patterns, e.g. fingerprints
62	Methods or arrangements for recognition using electronic means
64	using simultaneous comparisons or correlations of the image signals with a plurality of references, e.g. resistor matrix
66	references adjustable by an adaptive method, e.g. learning

G	PHYSICS
06	COMPUTING; CALCULATING OR COUNTING
K	RECOGNITION OF DATA; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
9	Methods or arrangements for reading or recognising printed or written characters or for recognising patterns, e.g. fingerprints
62	Methods or arrangements for recognition using electronic means
64	using simultaneous comparisons or correlations of the image signals with a plurality of references, e.g. resistor matrix
66	references adjustable by an adaptive method, e.g. learning

G	PHYSICS
06	COMPUTING; CALCULATING; COUNTING
N	COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
20	Machine learning
20	Ensemble learning

G	PHYSICS
06	COMPUTING; CALCULATING; COUNTING
N	COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
3	Computer systems based on biological models
02	using neural network models
08	Learning methods

G	PHYSICS
06	COMPUTING; CALCULATING; COUNTING
T	IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
17	Three dimensional [3D] modelling, e.g. data description of 3D objects
20	Finite element generation, e.g. wire-frame surface description, ; tesselation
205	Re-meshing

Processing

National Phase Entries

Authority File

Settings

Feedback

Goto Application

1. WO2023277975 - SYSTEMS AND METHODS OF CONTRASTIVE POINT COMPLETION WITH FINE-TO-COARSE REFINEMENT

G	PHYSICS
06	COMPUTING; CALCULATING; COUNTING
T	IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
2219	Indexing scheme for manipulating 3D models or images for computer graphics
20	Indexing scheme for editing of 3D models
2016	Rotation, translation, scaling

G	PHYSICS
06	COMPUTING; CALCULATING; COUNTING
T	IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
5	Image enhancement or restoration
001	Image restoration
002	Denoising; Smoothing