Saturday, 01 December 16:15 - 18:00 | Peridot 206
|
User-guided White Balance for Mixed Lighting ConditionsWe present a practical method for high-quality white balancing in scenes with complex lighting based on user-provided scribbles. It relies on what is most intuitive to humans, relative reflectance properties. Our system enables users to produce compelling results on a broad range of inputs, even for difficult configurations. Ivaylo Boyadzhiev, Cornell University |
|
Calibrated Image Appearance ReproductionWe present a fully calibrated model for reproducing the appearance of images under a wide range of viewing conditions and display characteristics. We base our model on human visual perception and combine aspects of tone reproduction, color appearance modelling and lightness perception to achieve accurate and visually pleasing reproduction. Erik Reinhard, Max Planck Institute for Informatics |
|
Coherent Intrinsic Images from Photo CollectionsWe present a method to decompose a photo collection of a scene into its intrinsic images components. We estimate for each image its illumination and a reflectance that is coherent across all views. Our method facilitates reflectance editing, and enables new applications such as transferring lighting across photographs. Pierre-Yves Laffont, REVES/INRIA Sophia-Antipolis |
|
Robust patch-based HDR reconstruction of dynamic scenesIn this paper, we propose a new approach to HDR reconstruction based on a novel energy-minimization formulation called the HDR image synthesis equation. This results in an patch-based algorithm integrates alignment and reconstruction in a joint optimization, making it more robust to camera/scene motion than previous techniques. Pradeep Sen, University of California, Santa Barbara |