Opponent Color Revisited
Sabine Süsstrunk, IVRG, School of Computer and Communication Sciences at EPFL
According to the efficient coding hypothesis, the goal of the visual system should be to encode the information presented to the retina with as little redundancy as possible. From a signal processing point of view, the first step in removing redundancy is de-correlation, which removes the second order dependencies in the signal. This principle was explored in the context of trichromatic vision by Buchsbaum and Gottschalk and later Ruderman et al., who found that linear de-correlation of the LMS cone responses matches the opponent color coding in the human visual system.
And yet, there is comparatively little research in computational photography and computer vision that explicitly model and incorporate color opponency into solving imaging tasks. A common perception is that "colors" are redundant and/or too correlated to be of any interest, or that it is too complex to deal with.
In this talk I will illustrate with several applications, such as saliency and super-pixels, that considering opponent colors can significantly improve computational photography and computer vision tasks not only in image enhancement but also image ranking. We have additionally extended the concept of "color opponency" to include near-infrared for applications such as scene recognition, object segmentation, and semantic image labeling.
Prof. Dr. Sabine Süsstrunk is a professor in the School of Information and Communication Sciences (IC) at the Ecole Polytechnique Fédérale de Lausanne (EPFL), Switzerland, where she leads the Images and Visual Representation Group since 1999. Her research areas are in computational photography, color computer vision and image processing, image quality and computational aesthetics. She has published over 120 scientific papers in peer-reviewed journals and conferences and holds 7 patents. She served as chair or committee member in numerous international conferences. She is involved in several spin-offs such as Croppola, a popular on-line tool and mobile app that intelligently crops a photo based on image composition. Sight.io, a photo-ranking engine that automatically sorts photographs on the basis of their aesthetic beauty and composition was recently acquired by EyeEm, a Berlin based community and market place for a new generation of photographers. In 2013, she received the IS&T/SPIE Electronic Imaging Scientist of the Year Award for her contributions to color imaging, computational photography, and image quality. She is a Fellow of IS&T.
Computational Imaging and Display – Hardware-Software Co-design for Imaging Devices
Wolfgang Heidrich, VCC – King Abdullah University of Science and Technology (KAUST) and University of British Columbia
Computational Imaging aims to develop new cameras and imaging modalities that optically encode information about the real world in such a way that it can be captured by image sensors. The resulting images represent detailed information such as scene geometry, motion of solids and liquids, multi-spectral information, or high contrast (high dynamic range), which can then be computationally decoded using inverse methods, machine learning, and numerical optimization. Computational Displays use a similar approach, but in reverse. Here, the goal is to computationally encode a target image that is then optically decoded by the display hardware for presentation to a human observer. Computational displays are capable of generating glasses-free 3D displays, high dynamic range imagery, or images and videos with spatial and/or temporal super-resolution. In this talk I will give an overview of recent advances and current challenges in rapidly expanding research area.
Prof. Wolfgang Heidrich is the director of the Visual Computing Center at King Abdullah University of Science and Technology (KAUST). He is also affiliated with the University of British Columbia, where he held the Dolby Research Chair until 2013. Dr. Heidrich received his PhD in Computer Science from the University of Erlangen in 1999, and then worked as a Research Associate in the Computer Graphics Group of the Max-Planck-Institute for Computer Science in Saarbrucken, Germany, before joining UBC in 2000. Dr. Heidrich's research interests lie at the intersection of computer graphics, computer vision, imaging, and optics. In particular, he has worked on computational photography and displays, High Dynamic Range imaging and display, image-based modeling, measuring, and rendering, geometry acquisition, GPU-based rendering, and global illumination. Dr. Heidrich has written well over 150 refereed publications on these subjects and has served on numerous program committees. His work on High Dynamic Range Displays served as the basis for the technology behind Brightside Technologies, which was acquired by Dolby in 2007 Dr. Heidrich has served as the program co-chair for Graphics Hardware 2002, Graphics Interface 2004, the Eurographics Symposium on Rendering, 2006, and PROCAMS 2011. Dr. Heidrich is the recipient of a 2014 Humboldt Research Award.
Open Problems and Current Directions in Physically-Based Rendering
Jaakko Lehtinen, NVIDIA Research and Aalto University School of Science
Realistic image synthesis (rendering) has been an integral part of computer graphics since its inception. Several luminaries have given us equations and algorithms that are now well-established, allowing applications to visualize worlds both real and unreal in a physically-based manner. However, current general-purpose algorithms are generally brittle in that they require content prepared by experts for reliable operation. On the other hand, the commoditization of large-scale 3D scanning and modeling techniques is quickly leading towards 3D models of entire cities, and, in time, the world. To resolve this conflict, I argue for the need for "output-sensitive" rendering algorithms. I will also give an overview of recent work in adaptive rendering, and provide my thoughts on other open problems in physically-based rendering as well.
I am an assistant professor at Aalto University, and a research scientist at NVIDIA Research. Prior to that, I did a postdoc with Frédo Durand at MIT CSAIL. I work mostly on realistic image synthesis, but my interests span most areas in graphics, including hardware architectures, modeling, mathematics, and appearance acquisition. I obtained my doctorate in 2007 from Helsinki University of Technology (now Aalto University), working together with Timo Aila, Janne Kontkanen and Samuli Laine under Lauri Savioja. In addition to research, I also consult for the game developer Remedy Entertainment, where I worked 1996-2005 as a graphics programmer and contributed significantly to the graphics technology behind Max Payne, Max Payne 2, and Alan Wake. In my free time, I enjoy rock climbing, photography, cooking, and music.