SHORT Papers

Short Papers selected for Poster Presentation

1. ADAPTIVE OBJECT PLACEMENT FOR AUGMENTED REALITY USE IN DRIVER ASSISTANCE SYSTEMS by Lucie Bordes, Toby P. Breckon, Alireza Kheyrollahi (Cranfield University, UK.)

Abstract
We present an approach for adaptive object placement for Augmented Reality (AR) use in driver assistance systems. Combined vanishing point and road surface detection enable the real-time adaptive emplacement of AR objects within a drivers’ natural field of view for on-road information display. This work combines both automotive vision and multimedia production aspects of real-time visual engineering.

2. OPTIMIZING LEAF DISTRIBUTION USING A REFERENCE PHOTO by Chuan Li and Peter Hall (University of Bath, UK)

Abstract
Image based tree modeling methods focus on generating the branch structure. Leaves are added to the model in a separate process using either botanic rules for high realism, or random sampling for computation efficiency. However, flaws in leave density distribution may result in a significantly different appearance compared to the reference image, somtimes unrealistic, even when the underlying branch is correct. We present a simple but efficient algorithm that generates leaves which match the reference image.

3. MEDIA TOUCH SURFACE FOR LIVE BROADCAST VIEWER SUBMITTED CONTENT INTERACTION by Lee Griffiths (The University of Salford, Manchester, UK) , Phil Smith (BBC North West, Quay House, Media City UK, Salford )

Abstract
The BBC deliver several live studio based news broadcasts which involve the presenters interacting with media such as pictures, video and other data whilst a discussion develops. In reality most of the interaction is rehearsed, and controlled by the show director whilst the broadcast is being delivered. The work presented here discusses the development and studio trial of an interactive surface based application which allows presenters to interact with viewer submitted media in real time. The authors intend to demonstrate the system.

Supplementary material can be found here

4. STEREOSCOPIC FOCUS MISMATCH MONITORING by Sergi Pujades, Frédéric Devernay (Inria Grenoble, France)

Abstract
We detect focus mismatch between views of a stereoscopic pair. First, we compute a dense disparity map. Then, we use a measure to compare focus in both images. Finally we use robust statistics to find which images zones have different focus. We show the results on the original images.

5. RECOVERING ARTICULATED POSE OF 3D POINT CLOUDS by Philipp Fechteler and Peter Eisert (Fraunhofer HHI / Humboldt University, Berlin, Germany)

Abstract
We present an efficient optimization method to determine the 3D pose of a human from a point cloud. The well-known ICP algorithm is adapted to fit a generic articulated template to the 3D points. Each iteration jointly refines the parameters for rigid alignment, uniform scale as well as all joint angles. In experimental results we demonstrate the effectiveness of this computationally efficient approach.

6. MULTIPLE VIEW SEGMENTATION AND MATTING by Markus Kettern, David C. Schneider and Peter Eisert (Fraunhofer Heinrich Hertz Institute, Berlin, Germany)

Abstract
We propose a robust and fully automatic method for extracting a highly detailed transparency-preserving segmentation of a person’s head from multiple view recordings, including a background plate for each view. At first, trimaps containing a rough segmentation into foregorund, background and unknown image regions are extracted exploiting the visual hull of an initial foreground-background segmentation. The background plates are adapted to the lighting conditions of the recordings and the trimaps are used to initialize a state-of-the-art matting method adapted to a setup with precise background plates available. From the alpha matte, foreground colours are inferred for realistic rendering of the recordings onto novel backgrounds.

7. STATISTICALLY USER CONTROLLED TEXTURE MIXING by D. Pickup, D. Cosker, P. Hall, P.Willis (University of Bath,UK)

Abstract
We propose an extension to an existing patch-based texture synthesis method allowing the user to intuitively control the mixing of patches from two or more example textures by defining a probability map which influences which input example to copy a patch from at each output location.

8. WEB-BASED EXPLORATION OF MULTI-VIEW CONTENT by Wolfgang Weiss, Werner Bailer, Christian Schober, Georg Thallinger (DIGITAL – Institute for ICT, JOANNEUM RESEARCH, Graz, Austria)

Abstract
We present a light-weight Web client for exploring multi-view content sets in post-production.

9. EMOTE, A NEW WAY OF CREATING ANIMATED MESSAGES FOR WEB ENABLED DEVICES by V. Helzle, S. Spielmann, N. Zweiling (Institute of Animation, Filmakademie Baden-Wuerttemberg)

Abstract
We present is a new way of creating superior animated messages for mobile devices. Using text and emoticons an individual animated short message can be created and shared with friends worldwide.

10. FREE-VIEWPOINT RENDERING FOR OUTDOOR SKIING by P.S. Schübel, R. Hodgson, O. Grau (BBC Research & Development, United Kingdom)

Abstract
This short paper demonstrates use of Free-viewpoint rendering in an outside broadcast production. For automatic 3D reconstruction from multiple cameras a shape-from-silhouette approach is used. Camera calibration is achieved with a new LED-based wand suitable for in- and outdoor use.

11. REAL TIME DEPTH ESTIMATION USING LINE RECURSIVE MATCHING by C. Riechert, F. Zilly, P. Kauff (Fraunhofer Institute for Telecommunications, Berlin, Germany)

Abstract
Depth Image Based Rendering (DIBR) is a key technology needed to create content for auto-stereoscopic displays and to adapt stereoscopic content to different screens sizes. Pixel-dense depth maps for stereoscopic videos are required for this purpose. In this paper we present a new depth estimator which generates HD resolution depth maps suitable for DIBR in a real-time environment.

12. HIGH PRECISION TOF-GUIDED DEPTH FROM STEREO FOR ROOM SCANNING by R. Nair, D. Kondermann (HCI, University of Heidelberg, Germany)

Abstract
We present our approach to reconstructing high quality depth maps from stereo and time of flight (TOF) images. Unlike many regularization methods, our system is physically
plausible. We employ a baseline block matching algorithm that is guided by the TOF data. First we reduce the disparity search space by only probing compatible values. Furthermore, the matching window shape is deformed based on a superpixel segmentation of the TOF depth image, resulting in more accurate matching results

13. INPAINTING WITH DEPTH FOR FILM POST-PRODUCTION by Yotam Doron (University College London), Jan Kautz (University College London), Jonathan Starck (The Foundry)

Abstract

We present a method for video inpainting in scenes featuring dynamic camera movement. Our technique can be used for rig removal, clean plate generation, and 2D to 3D conversion in scenes containing significant scaling and perspective changes.

14. SMOOTH TRANSITIONS FOR CONCATENATIVE MOTION SYNTHESIS by Peng Huang, Adrian Hilton (Centre for Vision, Speech and Signal Processing, University of Surrey, UK,)

Abstract
Previous concatenative motion synthesis/animation methods link different motion sequences together without any smoothing at transitions. This often results in artefacts - sudden changes in both geometry and appearance. This paper provides a simple way to obtain smooth transitions by linear blending of temporal aligned 3D video frames both in geometry and appearance. The transition artefacts are appreciably reduced.

15. VIDEO RE-SAMPLING AND CONTENT RE-TARGETING FOR REALISTIC DRIVING INCIDENT SIMULATION by Alejo M. Heras, Toby P. Breckon, Marko Tirovic (School of Engineering, Cranfield University, UK.)

Abstract
We present a video generation and visualization system for integration into a driving incident simulator. We combine the use of video re-sampling, target tracking and object segmentation for content re-targeting to generate realistic video content that can be adapted on demand, based upon simulator inputs to generate a range of driver incident test scenarios