| Channel: eurographics 2011 |
Your post is to long. Please make it short.
new posts available!
| The speaker is from Japan and started his talk with a great "Thank you to all nations which supports Japan people after the natural disaster". | |||
|
| Last talk in this session is: "An Eyeglass Simulator using Conoid Tracing". | |||
|
| The method is about to render binary images, so only black and white color. The authors define this problem as energy minimization problem and use output from a rendering pipeline (e.g. diffuse colors, specular component) to setup the minimization terms. | |||
|
| Next paper is Binary Shading using Geometry Appearance. | |||
|
| Authors define the mosaicing problem as an energy minimization problem by giving costs like visibility, orientation, etc. to each tile. Finally it is optimized with graph cut. | |||
|
| Today I am in the Computer Graphics Forum Paper session. The first paper presented here is Generating Classic Mosaics with Graph Cuts. | |||
|
| This was the last paper session for today. Next, today at 19:00, there will be a welcome reception as also a poster session combined. Can't wait to see for the interesting posters and maybe taste food :) | |||
|
| I didn't quit get the main contribution of this paper, so can't really summarize it. In general the authors tries to utilize the GPU more when doing hybrid bidirectional path-tracing rendering. | |||
|
| Last talk in this session: "Combinatorial Bidirectional Path-Tracing for Efficient Hybrid CPU/GPU Rendering" | |||
|
| Javor presents an acceleration structure for ray tracing which is an extension of the uniform grid made by additional level: every grid node contains another full uniform grid. The hard part of their work is a massive parallel method for its construction. | |||
|
| Next talk by Javor Kalojanov "Two-Level Grids for Ray Tracing on GPUs" | |||
|
| In order to improve performance for occlusion rays, authors do extend surface area heuristic (SAH) to do ray termination (RTSAH), by replacing several assumptions stated for SAH. This improves for example Kd-Tree building. | |||
|
| Ok, after a short #coffee break, I've moved to "Ray Tracing" session. First talk in this session is "RTSAH Traversal Order for Occlusion Rays". | |||
|
| The authors use some aspect from perceptual laws, in order to correctly combine and separate symmetric parts into a hierarchical representation. The shape is first separated into parts by "concavity cuts" and refined by symmetry detection. | |||
|
| Last talk of this session: Symmetry Hiearchy of Man-Made Objects. | |||
|
| @at who is there? | |||
|
| Michael is presenting an approach capable of detecting symmetric parts in a 3D shape fully automatically. Hereby the detected parts are not exactly similar, they can be deformed. The approach can learn covariance from that found parts. | |||
|
| Based on the questions stated to Martin one could see that the audience is full of "Shape"-celebreties, so I am already scared for my talk :) | |||
|
| Next talk by Michael Wand: Shape Analysis with Subspace Symmetries. | |||
|
| @Alexander Berner, I can hardly see, since I am in the first row ;) | |||
|
| "Learning Line Features..." - an approach to find reoccurence of user selected point chain in scanned data. The authors use MRF chain and compute a solution for it by Max-Product Belief Propagation. | |||
|
| @at many people at the shape analysis session? Any celebrities? ;) | |||
|
| @Alexander Berner, was pretty good. Masu went the cool way, by not saying anything and let slides to explain everything. Michael managed to explain the full approach in 40sec, respect! :) | |||
|
| @at ah, time shift to uk :) | |||
|
| I am at the session Shape Analysis. The first talk is given by Martin Sunkel, Learning Line Features in 3D Geometry. | |||
|
| @at how were the talks, given by masu and michael? | |||
|
| cannot say exactly, seems to be that this year I wouldn't give any "most original" fast forward award to anybody | |||
|
| @at what was the best idea? | |||
|
| IMHO, in average fast forward talks given by people from our institue were the most creative and funniest. | |||
|