Tutorial 7: High dynamic range techniques in graphics: from acquisition to display
Tuesday, August 30th, 2005. 14:00 - 17:30 (Half Day)
VENUE: Emmet Theatre.
| ORGANISERS: |
Karol Myszkowski, MPI Informatik
Wolfgang Heidrich, University of British Columbia
|
| SPEAKERS: |
Michael Goesele (MPI)
Wolfgang Heidrich (UBC)
Bernd Höfflinger (IMS CHIPS)
Grzegorz Krawczyk (MPI)
Karol Myszkowski (MPI)
Matthew Trentacoste (UBC) |
Outline and Syllabus
This course is motivated by tremendous progress in the development and accessibility of high
dynamic range technology (HDR) that happened just recently, which creates many interesting
opportunities and challenges in graphics. The course presents a complete pipeline for HDR
image and video processing from acquisition, through compression and quality evaluation,
to display. Also, successful examples of the use of HDR technology in research setups and
industrial applications are provided. Whenever needed relevant background information on
human perception is given which enables better understanding of the design choices behind
the discussed algorithms and HDR equipment.
Description of images, from left to right:
- Acquisition of HDR data: two HDR cameras (IMS-CHIPS and LarsIII) and
a high quality LDR camera Jenoptik C14
-
HDR video: a broad range of luminance values can be stored in the HDR
video without loss of any perceivable details. Even the bare light bulb
is faithfully registered, as visible in the blue window which scales
down the luminance range.
-
HDR video: a conventional video in this shot could be exposed either
for the interior or exterior of the cafeteria. With the HDR video it is
possible to capture and store all details of the recorded scene.
-
Tone mapping: most display devices are still limited in the dynamic
range of luminance and contrast that can be reproduced (right inset
illustrates a naive display of an HDR image). Various tone mapping
techniques allow for reduction of such a contrast. Here the image is
decomposed into the areas of consistent illumination (left inset) and
the contrast ratio between these areas is optimized.
-
HDR display: a display system developped by the Sunnybrook
company that is capable of displaying images
with a dynamic range much more similar to that encountered in the real
world.
-
Introduction - Myszkowski (10 mins)
- HDR Acquisition Techniques for Still Images and Video -
Goesele and Krawczyk (35 mins)
- Overview over HDR Acquisition Techniques
- Multi-exposure Techniques
- HDR Cameras and their Photometric Calibration
- Absolute Luminance Calibration
- HDR Image Sensors for Video - Hoefflinger (20 mins)
- CMOS Active-pixel (APS)
- Extended Dynamic Range Sensor (XDR)
- HDRC Imager (Very High Dynamic Range of 170 dB)
- HDR Image and Video Compression - Myszkowski (30 mins)
- Luminance and Chroma Quantization Solutions
- HDR Image and Video Formats
- HDR Quality Metrics and Their Validation
- Real-time Post-processing of HDR Video Stream (Live demo)
- Break
- HDR Displays - Heidrich and Trentacoste (35 mins)
- Foundations: Limitations of Human Perception
- Hardware Design
- Processing Algorithms
- HDR Applications - Goesele, Krawczyk, and Hoefflinger (45 mins)
- Image-based Measurements of Object and Material Properties
- Virtual Scene Re-lighting: CAVE System for Car Interior Modeling
- HDR Lighting in Mixed-Reality Applications
- Application Examples in Automotive Industry and Computer Vision
- Final Remarks and Open Discussion (10 mins)
Tutorial Presenters' Information
Michael Goesele is a postdoctoral researcher in the computer graphics group of Prof. Hans-
Peter Seidel at the MPI Informatik in Saarbrcken, Germany. After receiving his diploma
in computer science from the University of Ulm in 1999 he joined the MPI Informatik and
received his PhD from Saarland University in 2004. His research is focused on acquisition
techniques for computer graphics making active use of HDR technology. Among others, he
recently published two papers at ACM SIGGRAPH about the acquisition of light sources
(Accurate Light Source Acquisition and Presentation) and translucent objects (DISCO
Acquisition of Translucent Objects). He has given several lectures and tutorials (e.g. at
Eurographics 2002) about digitization techniques and has several years of experience with HDR
imaging.
Wolfgang Heidrich is an Associate Professor in the Department of Computer Science at the
University of British Columbia. He received his PhD in Computer Science from the University
of Erlangen in 1999, and then worked as a Research Associate at the Computer Graphics
Group of the Max-Planck-Institute for Computer Science in Saarbrucken, Germany, before
joining UBC in 2000. Heidrich's research interests lie in computer graphics, computer vision,
and imaging. In particular, he has worked on image-based modeling, measuring, and rendering,
HDR imaging, geometry acquisition, hardware-accelerated and image-based rendering,
and global illumination. Heidrich has published over 60 peer-reviewed articles on these
subjects and has served on numerous program committees. He was the program co-chair for
Graphics Hardware 2002, and Graphics Interface 2004, and he is a co-author of the recently
published book Real-Time Shading.
Bernd Hoeflinger is Director of the Institute for Microelectronics Stuttgart (IMS),
a certified contract R&D and Manufacturing Foundation for application-specific microelectronic
systems. He has held R& D, faculty and leadership positions at Siemens and at Cornell
University as well as the Universities of Dortmund, California (Berkeley), Minnesota, Purdue
and Stuttgart. He has been a pioneer in CMOS technology, recently with a focus on CMOS
imagers and cameras. The automotive and manufacturing industry is the leading group of
customers and partners of his institute.
Grzegorz Krawczyk is a PhD student at the Max-Planck-Institute fr Informatik, Germany.
His principal areas of research include High Dynamic Range image and video processing, in
particular perception-inspired tone mapping operators and photometric calibration of High
Dynamic Range video cameras. He received his MSc degree from the Technical University
of Szczecin, Poland. His recent publications include Perception-motivated High Dynamic
Range Video Encoding in ACM Transactions on Graphics (Proceedings of ACM SIGGRAPH
2004) and A CAVE System for Interactive Modeling of Global Illumination in Car Interior
in Proceedings of ACM Symposium on Virtual Reality Software and Technology, 2004.
Rafal Mantiuk is a PhD student at the Max-Planck-Institute fr Informatik, Germany. His
principal areas of research include compression and processing of high-dynamic range video,
perception issues in computer graphics, and visual quality metrics. He received his MSc
degree from the Technical University of Szczecin, Poland. His recent publications include
Perception-motivated High Dynamic Range Video Encoding" in ACM Transactions on Graphics
(Proceedings of ACM SIGGRAPH 2004) and Predicting Visible Differences in High Dynamic
Range Images - Model and its Calibration" in Human Vision and Electronic Imaging
2005 (SPIE).
Karol Myszkowski is a senior researcher in the Computer Graphics Group of Max-Planck-
Institut fr Informatik. From 1993 to 2000 he served as a tenured Associate Professor at the
University of Aizu, Japan. In the period 1986-1992 he worked with Japanese company Integra,
Inc. developing rendering software for such customers as Toshiba Lighting, Shiseido,
Matsushita Electric, Kandenko, and others. He received his PhD and habilitation degrees
in computer science from Warsaw University of Technology (Poland) in 1991 and 2001, respectively.
His research interests include perception issues in graphics, high-dynamic range
imaging, global illumination and rendering. He was the Program Committee co-chair of Eurographics
Rendering Workshop, 2001. He has been a speaker for Eurographics STAR 2002
and SIGGRAPH courses in 2000, 2001, and 2003. His recent papers relevant to the scope
of this course have been published in the proceedings of Siggraph 2004, Eurographics 2003,
ACM VRST 2004, and Human Vision and Electronic Imaging 2003 and 2005 (SPIE).
|
|
