• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2020.11a
• /
• pp.166-168
• /
• 2020

MPEG 비디오 그룹은 MPEG-I 표준의 일부로 포인트 클라우드(Point Cloud) 압축을 위한 비디오 기반 포인트 클라우드 부호화(V-PCC)와 몰입형(immersive) 비디오 압축을 위한 MPEG Immersive Video(MIV) 표준을 개발하고 있다. 최근에는 포인트 클라우드 및 몰입형 비디오와 같은 체적형(volumetric) 비디오를 모두 압축할 수 있도록 V-PCC 와 MIV 를 통합한 V3C(Visual Volumetric Video-based Coding) 표준화를 진행하고 있다. 본 논문에서는 V3C 코덱을 사용한 3DoF+(3 Degree of Freedom plus) 비디오 부호화 방안을 분석한다. 또한 V3C 코덱의 2D 코덱으로 기존 HEVC 대신 VVC 를 사용할 경우의 부호화 성능 향상을 분석한다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2021.06a
• /
• pp.151-153
• /
• 2021

포인트 클라우드는 특정 개체 혹은 장면을 다수의 3 차원 포인터를 사용하여 표현하는 데이터의 표현 방식 중 하나로 3D 데이터를 정밀하게 수집하고 표현할 수 있는 방법이다. 하지만 방대한 양의 데이터를 필요로 하기 때문에 효율적인 압축이 필수적이다. 이에 따라 국제 표준화 단체인 Moving Picture Experts Group 에서는 포인트 클라우드 데이터의 효율적인 압축 방법 중 하나로 Video based Point Cloud Compression(V-PCC)에 대한 표준을 제정하였다. V-PCC 는 포인트 클라우드 정보를 Occupancy, Geometry, Texture 와 같은 다수의 2D 영상으로 변환하고 각 2D 영상을 전통적인 2D 비디오 코덱을 활용하여 압축하는 방법이다. 본 논문에서는 V-PCC 에서 변환하는 Occupancy 의 정보를 활용하여 효율적으로 Texture 영상을 압축할 수 있은 방법을 소개한다. 또한 제안방법이 V-PCC 에서 약 1%의 부호화 효율을 얻을 수 있음을 보인다.

• KIPS Transactions on Software and Data Engineering
• /
• v.12 no.12
• /
• pp.505-518
• /
• 2023

3D point cloud semantic segmentation is a computer vision task that involves dividing the point cloud into different objects and regions by predicting the class label of each point. Existing 3D semantic segmentation models have some limitations in performing sufficient fusion of multi-modal features while ensuring both characteristics of 2D visual features extracted from RGB images and 3D geometric features extracted from point cloud. Therefore, in this paper, we propose MMCA-Net, a novel 3D semantic segmentation model using 2D-3D multi-modal features. The proposed model effectively fuses two heterogeneous 2D visual features and 3D geometric features by using an intermediate fusion strategy and a multi-modal cross attention-based fusion operation. Also, the proposed model extracts context-rich 3D geometric features from input point cloud consisting of irregularly distributed points by adopting PTv2 as 3D geometric encoder. In this paper, we conducted both quantitative and qualitative experiments with the benchmark dataset, ScanNetv2 in order to analyze the performance of the proposed model. In terms of the metric mIoU, the proposed model showed a 9.2% performance improvement over the PTv2 model using only 3D geometric features, and a 12.12% performance improvement over the MVPNet model using 2D-3D multi-modal features. As a result, we proved the effectiveness and usefulness of the proposed model.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.32 no.2
• /
• pp.109-115
• /
• 2014

The aim of the study is to present a method to build an octree and to query from it for huge 3D point clouds of which volumes correspond or surpass the main memory, based on the memory-efficient octree developed by Han(2013). To the end, the method directly refers to 3D point cloud stored in a file on a hard disk drive instead of referring to that duplicated in the main memory. In addition, the method can save time to rebuild octree by storing and restoring it from a file. The memory-referring method and the present file-referring one are analyzed using a dataset composed of 18 million points surveyed in a tunnel. In results, the memory-referring method enormously exceeded the speed of the file-referring one when generating octree and querying points. Meanwhile, it is remarkable that a still bigger dataset composed of over 300 million points could be queried by the file-referring method, which would not be possible by the memory-referring one, though an optimal octree destination level could not be reached. Furthermore, the octree rebuilding method proved itself to be very efficient by diminishing the restoration time to about 3% of the generation time.

• Broadcasting and Media Magazine
• /
• v.26 no.2
• /
• pp.31-45
• /
• 2021

본 고는 ISO/IEC JTC 1/SC 29/WG 7 MPEG(Moving Picture Experts Group) 3DG(3D Graphics coding) 그룹에서 진행되고 있는 포인트 클라우드 데이터 압축 표준 기술 중 하나인 G-PCC(Geometry-based Point Cloud Compression) 표준에 대하여 설명하고자 한다. G-PCC는 포인트 클라우드의 기하 정보와 속성 정보를 3차원 공간에서 서로 다른 기술을 이용하여 압축하는 표준으로, 무손실 압축 방법의 경우 10:1의 압축율을 제공하고 손실 압축의 경우 35:1 정도의 압축율을 보인다. 본 고에서는 G-PCC의 기하 정보와 속성 정보의 압축 방법을 상세히 설명하고 같은 기능을 수행하는 압축 기술 간의 성능을 비교하고자 한다.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2022.06a
• /
• pp.1275-1277
• /
• 2022

Structure-from-Motion(SfM), Multi-view Stereo(MVS)이 이용되는 3D Reconstruction 과정에서 생성된 3D 포인트 클라우드는 RGB 영상에 기반하여 생성되므로 실제 객체 혹은 Scene 과 달리 point 와 point 간에 존재하는 빈 공간이 발생한다. 이를 개선하기 위하여 3D 포인트 클라우드를 이용하여 3D Mesh 를 복원하는 Mesh Reconstruction 과정을 거치게 된다. 본 논문에서는 Mesh Reconstruction 과정에서 자유공간 지지도에 기반해 수정한 가중치를 이용하는 효율적인 방법을 제안한다. 실험을 통하여 제안한 알고리즘을 이용한 복원 결과가 기존보다 개선됨을 보인다.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.19 no.5
• /
• pp.1199-1206
• /
• 2015

This paper introduces an algorithm for computing similarity between two poses in the motion capture data with different scale of skeleton, different number of joints and different joint names. The proposed algorithm first performs the topological analysis on the skeleton hierarchy for classifying the joints into more meaningful groups. The global joints positions of each joint group then are aggregated into a point cloud. The number of joints and their positions are automatically adjusted in this process. Once we have two point clouds, the algorithm finds an optimal 2D transform matrix that transforms one point cloud to the other as closely as possible. Then, the similarity can be obtained by summing up all distance values between two points clouds after applying the 2D transform matrix. After some experiment, we found that the proposed algorithm is able to compute the similarity between two poses regardless of their scale, joint name and the number of joints.

• Journal of KIISE
• /
• v.43 no.5
• /
• pp.562-568
• /
• 2016

Point cloud data can be expressed in a specific coordinate system of a data set with a large number of points, to represent any form that generally has different characteristics in the three-dimensional coordinate space. This paper is aimed at finding a cylindrical pipe in the point cloud of the three-dimensional coordinate system using RANSAC, which is faster than the conventional Hough Transform method. In this study, the proposed cylindrical pipe is estimated by combining the results of parameters based on two mathematical models. The two kinds of mathematical models include a sphere and line, searching the sphere center point and radius in the cylinder, and detecting the cylinder with straightening of center. This method can match cylindrical pipe with relative accuracy; furthermore, the process is rapid except for normal estimation and segmentation. Quick cylinders matching could benefit from laser scanning and reverse engineering construction sectors that require pipe real-time estimates.

• Cartoon and Animation Studies
• /
• s.48
• /
• pp.199-215
• /
• 2017

Recently, a drone is used for the general purpose application although the drone was builtfor the military purpose. A drone is actively used for the creation of contents, and an image acquisition. In this paper, we develop a 3D library module platform using 3D mesh model data, which is generated by a drone image and its point cloud. First, a lot of 2D image data are taken by a drone, and a point cloud data is generated from 2D drone images. A 3D mesh data is acquired from point cloud data. Then, we develop a service library platform using a transformed 3D data for multi-purpose uses. Our platform with 3D data can minimize the cost and time of contents creation for special effects during the production of a movie, drama, or documentary. Our platform can contribute the creation of experts for the digital contents production in the field of a realistic media, a special image, and exhibitions.

• Journal of Korean Society for Geospatial Information Science
• /
• v.20 no.3
• /
• pp.3-10
• /
• 2012

The investigation of valley networks and their volume provide important information about past water activities on Mars. As an alternative of conventional image processing methods, terrain filtering algorithm using pointcloud data is suggested in this study. First, the topography of pointcloud is inverted so that the valleys become positive features and the algorithm is then applied to distinguish the valleys from the surface. Ground DEM and object DEM are generated from both the valleys and the surface pointcloud then the volume of valleys is estimated by multiplying the height difference between the surface with valleys and the area of valleys based on grid cellsize. In the test of valleys adjacent to Tuscaloosa crater, the total volume of valleys was estimated to be $1.41{\times}10^{11}m^3$ with the difference of 12% and 16% compared with the infill volume of Tuscaloosa crater and BTH result respectively.