제한없는 임의 시공간에서 취득한 영상으로부터 완전한 얼굴 복원
Total Reconstruction of Faces from Images Captured under Unconstrained and Arbitrary Space-Time
(2019.03 ~ 2024.02)

• 본 과제에서는 제한되지 않은 임의의 시공간 환경에서 취득된 일반적인 얼굴 영상으로부터의 완전한 얼굴 정보를 복원하고 미래의 정보를 예측하는 것을 목표로 함

• Funded by National Research Foundation (한국연구재단 중견연구자지원사업)

다중 스테레오 카메라 보정 및 깊이영상 추정기법 개발
Multi-view stereo camera calibration and depth estimation
(2022.07 ~ 2023.12)

• 본 과제에서는 다중 스테레오 카메라의 보정과 정밀한 깊이 영상 취득을 목표로 함

• Funded by Vatech (주식회사 바텍 산학협력과제)

시공간 제한 없는 비디오에서 다중 휴먼 3차원 형상, 자세, 동작 복원 및 예측
Holistic Multi-Human Reconstruction and Prediction from Videos with Unconstrained Space-Time
(2019.06 ~ 2022.09)

• CCTV, 스마트폰, 유튜브 비디오, 고전 영화 등 제한 없는 환경에서 취득한 일상생활의 임의의 비디오에 존재하는 다수의 휴먼 객체의 형상(shape), 자세(pose), 동작(motion)을 시공간 일관성 있게 정확히 추정하고 심층신경망 학습을 통해 미래의 동작과 비디오를 예측

• Funded by Samsung Research Funding & Incubation Center for Future Technology (삼성미래기술육성사업)

스마트기기를 위한 온디바이스 지능형 정보처리 가속화 SW플랫폼 기술 개발
Development of Acceleration SW Platform Technology for On-device Intelligent Information Processing on Smart Devices
(2017.03 ~ 2021.12)

• Embedded GPU를 장착한 스마트 기기에서 양안 스테레오, 다중뷰 스테레오, 비디오 등의 입력 영상으로부터 3차원 장면 정보를 고속 추출하고 이를 모바일 증강현실에 응용하기 위한 OpenCL 라이브러리 구축

• Funded by Institute for Information & communications Technology Promotion (IITP) (정보통신기획평가원 정보통신·방송연구개발사업, 주관연구기관: 한국전자통신연구원)

• 온디바이스 공간 정보처리 가속 라이브러리 cross reference (양안 스테레오 정합 / 멀티뷰 스테레오 정합 / 라이트필드 생성 및 정합 / 신경망 기반 멀티뷰 깊이 추정 / 차폐인지 증강현실 어플리케이션)

     
    Library : https://etri-ivcl-document.netlify.app/
    Video : https://youtu.be/xDCV1u3dY30 

(기술 이전을 희망하는 기업은 032-860-9190로 연락 부탁드립니다)

라이트 필드 기반의 초시공간 360도 영상 취득 및 처리 기술
Light Field based Time-Space Transcendence 360 Video Reconstruction and Processing
(2017.12 ~ 2020.11)

• 다양한 종류의 2차원 영상 데이터로부터 광대한 스케일의 시간(1차원)과 공간(3차원)을 자유롭게 거닐 수 있는 4차원 Light Field 영상 데이터를 구성하여 증강현실 및 가상세계에서 사실적인 시각적 재현이 가능한 기술을 확보한다.

• Funded by Samsung Research Funding & Incubation Center for Future Technology

Computational Image Reconstruction on 4D Light Field
(2016.06 ~ 2019.05)

• We will address the problem of computational image reconstruction (deblurring, layer separation, superresolution, dehazing, etc) on 4D light fields.

• Funded by National Research Foundation (미래창조과학부 중견연구자지원사업)

Image-Based Human Modeling for Augmented Reality
(2017.03 ~ 2018.04) 

• Image-based face, body, hair modeling for AR and VR

• Funded by CoAsia Korea

Mobile Augmented Reality
(2016.07 ~ 2017.06)

• Video augmented reality (AR) based on RGBD visual SLAM

• Funded by SK Telecom

Light Field Image Acquisition, Processing, and Understanding
(2013.12 ~ 2016.11)

• Developing 4D light field acquisition and processing algorithms which overcome the limitation of conventional imaging

• Funded by National Research Foundation (미래창조과학부 중견연구자지원사업)

3D Scene Reconstruction Using Car Blackbox Images
(2013.9 ~ 2016.8)

• 3D scene reconstruction using car blackbox images for analyzing car accident

• Funded by The Ministry of Science, ICT and Future Planning (미래창조과학부 디지털콘텐츠원천기술개발사업)

Mobile CPU-GPU Fusion Processor
(2012.6 ~ 2015.5) 

• Developing image processing framework on handheld GPU using OpenGL ES

• Funded by The Ministry of Knowledge Economy (지식경제부 산업융합원천기술개발사업)

GPU Parallel Energy Minimization on Massive Graphical Models
(2012.5 ~ 2015.4) 

• Global optimization using graph cuts using massive GPU parallel computing and its application to 3D TV (stereo images)

• Funded by National Research Foundation (교육과학기술부 일반연구자지원사업)

Low-Power Superresolution on Embedded GPU
(2013.07 ~ 2014.07) 

• Funded by LG Electronics (LG전자)

Content-Based 3D Model Retrieval in Mobile Environment
(2012.10 ~ 2013.6)

• Developing 3D shape retrieval algorithm using noisy and partial 3D depth image

• Funded by The Ministry of Knowledge Economy (지식경제부) and NHN Corporation

Image Quality Enhancement for Smartphone
(2013.5 ~ 2013.10) 

• Developing fast and robust image processing algorithms for Galaxy S and Galaxy Note series

• Funded by Samsung Electronics (삼성전자)

Real-Time Feature Extraction and Tracking on the GPU
(2012.5 ~ 2012.8)

• Real time SURF feature extraction and tracking (for shape from motion) using CUDA on the GPU

• Funded by Electronics and Telecommunication Research Institute (ETRI)

Media Retargeting for Stereoscopic Video
(2011.10 ~ 2012.8)

• Developing content-adaptive image resizing algorithm for stereoscopic images and videos

• Funded by The Ministry of Knowledge Economy (지식경제부) and LG Electronics (LG전자)

Multiview Inverse Motion Photography
(2009.5 ~ 2012.6)

• Multiview motion deblurring and its application to 3D reconstruction

• Funded by National Research Foundation (교육과학기술부 기초연구과제)

Video Summarization and 3D Effect for Smartphone
(2011.4 ~ 2012.1)

• Developing video summarizatioin and 3D transition algorithms for brand-new Galaxy S and Galaxy Tab

•  Funded by Samsung Electronics (삼성전자)

Realistic Broadcasting Research Center
(2009.1~2011.12)

• 3D video acquisition, transmission, and rendering

• Funded by The Ministry of Knowledge Economy (대학IT연구센터(ITRC))

High Dynamic Range Video
(2008.9 ~ 2011.8)

• Interdisciplinary research of image processing, vision, communication technologies for acquisition and processing of high dynamic range video

• Funded by Korea Science and Engineering Foundation (한국연구재단 특정기초연구지원사업)

Machine Vision Inspection Using GPU Parallel Computing
(2011.04 ~ 2011.8)

• Parallelizing LCD panel inspection algorithm on the GPU

• Funded by LG Electronics (LG전자)

3D Shape Acquisition on a Smartphone
(2010.12 ~ 2011.7)

• Developing photometric stereo algorithm using mobile phone's camera and flash

• Funded by The Ministry of Knowledge Economy (지식경제부) and Samsung Electronics (삼성전자)

Accelerating Computer Vision Algorithms on Handheld GPU
(2010.7 ~ 2010.12)

• Developing stereo matching and feature detection algorithms on PowerVR SGX 540 for Samsung Galaxy S

• Funded by Samsung Electronics (삼성전자)

Automatic 3D face modeling on a smartphone
(2010.4 ~ 2010.11)

• Developing 3D face modeling algorithm using 2 views captured by and processed on a Samsung's Bada smartphone

• Funded by Samsung Electronics (삼성전자)

3D Shape Reconstruction by Multi-View Stereo (MVS)
(2008.9~2010.2)

• Multiresolution volumetric/surfel modeling and rending using multiple views on the GPU

• Funded by Ministry of Culture, Sports and Tourism (IT 원천기술개발사업)

Computational Photography (Mobile Photo Collage)
(2009.5 ~ 2009.11)

• Automatic collage generation on the mobile handset (Samsung Omnia II)

• Funded by Samsung Electronics (삼성전자)

Computational Motion Photography
(2008.7~2009.6)

• Interactive motion depiction using efficient user interface

• Funded by Korean Research Foundation (학술진흥재단 신진교수연구과제)

3D Pose Estimation on the GPU
(2007.1 ~ 2008.2)

• Jointly Funded by Inha University, Korea Research Foundation (학술진흥재단 국제공동연구지원), and Mitsubishi Electric Research Laboratories (MERL)

• Performed during 1-year visiting research at MERL

Digital actor and 3-D avatar creation
(2005.5 ~ 2007.4)

• Automatic and UI-assisted parametric 3-D face modeling and animation

• Funded by Korean Ministry of Commerce, Industry, and Energy (산업자원부 지역산업 기술혁신사업)

GPU-based image processing library
(2006.3 ~ 2006.11)

• Design and development of image processing framework based on GPU and programmable shader

• Funded by Samsung Electronics (삼성전자 위탁연구과제)

Multiresolution processing of 3-D video
(2005.7~2006.6)

• Multiresolution rendering of image-based and model-based 3-D video sequence

• Funded by Korea Research Foundation (학술진흥재단 신진교수연구과제)

Image-based 3-D surface modeling
(2005.6 ~ 2005.12)

• NURBS surface construction from matched curves in multiple calibrated images

• Funded by Korean Ministry of Information and Communication (전자통신연구원 위탁연구과제)

3-D shape description and database retrieval
(2002.7~2004.6)

• Collaboration with Samsung Advanced Institute of Technology and Seoul National University

• Proposed to and adopted by MPEG-7 Visual Ver.3

Image-based modeling and rendering
(2001~2004)

• Collaboration with Graphics and Media Lab. in Moscow State University

• Proposed and adopted as an international standard in MPEG-4 AFX (Animation Framework eXtension)

3-D photorealistic face modeling
(2002~2004)

• Collaboration with Graphics and Media Lab. in Moscow State University