First session of the day was a set of paper presentations all on "Image Based Modelling". Some very cool ideas here.The first was videograph which allows a user to extract a 3D model from video. It very easy to use, basically the user just draws rectangles (and other shapes) to define areas they want to extract (say the bonnet, doors, wheels etc of a car) and then in each frame of video they adjust the points to their correct positions or add new points as more of the object becomes visible. A simple click later and the software creates a textured 3D model that can be used in games or to add into other video clips.
The second paper covered the building of a 3D model of a building just from a picture of its facade. It also was able to generate a grammar describing the building so a procedural engine could be used to build new buildings in the same style. This was done by some of the guys who did the urban design and procedural modelling course, on day 2, that I attended.
The last two papers both covered the generation of accurate tree models from photographs. It was interesting to see the different approaches taken and how they affected the final tools.
The last session of the day was a "sketches" session in which quick overviews of technologies or techniques are given. The one I went to was titled "Drat more rats" and was given by the Pixar guys about all the new technology they used to do their new Ratatouille movie. Some very interesting soft body dynamics and fluid simulations going on here. Hopefully this movie's story line is good because the visuals are fantastic.It was really interesting to see how much simulation is used to generate today's 3D animations but how the holy grail seems to be to still give artists control of the process so that the simulations can be direct for arts sake.
No comments:
Post a Comment