Using Space Effectively: 3D
From Visualization Sp06
Lecture on Mar 7, 2006
Readings
- Chapter 2, Projection Systems. In Art and Representation. Willats (handout)
- Chapter 8: Marginal Distortions. In The Psychology of Perspective and Renaissance Art. Kubovy (pdf)
- What object attributes determine canonical views? Blanz et al. (html) (pdf)
Optional Readings
- Artistic Multiprojection Rendering. Agrawala et al. (html)
- Automatic View Selection Using Viewpoint Entropy and its Application to Image-Based Modelling. Vázquez et al. (html)
- Artistic Composition for Image Creation. Gooch et al. (pdf)
- Map Projections in PDF
Pushkar - Mar 07, 2006 11:58:03 am
Since a lot of the readings talk about perspective projections for art (which started showing up only after the Renaissance), I thought I'd mention this additional link:
http://www.handprint.com/HP/WCL/tech10.html
It contains some useful info. on the type of perspective projections.
Lesliei - Mar 07, 2006 05:34:00 pm
Rudolf Arnheim (Professor Emeritus of the Psychology of Art at Harvard) in "Visual Thinking" talks about the perspective view, and the actively-looking, moving eye versus the passive gaze. Unfortunately, I can't find a link to any parts of this book, but I'm sure the library has it. Alternatively, if you send me an e-mail (lesliei@cs), I have a copy of some parts.
Bryan - Mar 09, 2006 09:36:32 am
I mentioned the Robinson projection as a better alternative to choosing only the preservation of angle or size. [Here] is a link to wikipedia's discussion of the projection with some nice pictures. The Robinson projection used to be the official projection for National Geographic, but it was recently replaced by [the Winkel Tripel], which is actually older.
Sychan - Mar 09, 2006 12:53:43 pm
I was thinking about the different map views which were presented, some of which maintained correct angles, while others maintained correct area and the tradeoffs which needed to be made. But it seemed that many of those tradeoffs were due to the fact that it was a spherical surface being mapped onto a planar surface, and the scale of the map. Folks had to worry about these issues because the display medium was static.
How much do we have to worry about these kinds of tradeoffs, and what are the "wins" from the various approaches if we have interactive mediums that allow us to zoom in, change our relative angle, etc...? Are the more interesting problems these days figuring out how to compose different perspective approaches? I'm thinking specifically about the 3D display wall and hyperbolic fisheye visualizations that have some kind of focus + context approach. The focus region is kind of "flat", but the context region uses various forms of perspective to represent some kind of summarized context.
--Sychan 13:11, 9 March 2006 (PST)
Ashley - Mar 13, 2006 06:30:35 pm
There has been much speculation as to whether optical devices were used to help early Renaissance painters, and one Jan van Eyck painting in particular is often cited as evidence, including by Hockney, who has written books/documentaries on the topic. This interesting article describes work done recently by computer vision researchers that appears to debunk the claim. I personally think it's a really fascinating use of perspective/computer vision techniques, coupled with a little bit of historical information (e.g. how a curved mirror would have been made at the time, manufacturing tolerance/probable size of chandeliers from 15th century Europe, etc.) The article itself is very accessible and has lots of pictures.
http://www.osa-opn.org/view_file.cfm?doc=%24(LK.J0%20%20%0A&id=%24)L7%2BK%40%20%20%0A
Raymond - Mar 14, 2006 01:14:20 pm
Since this lecture is all about perspectives, I thought I'd post a link to some cool (aerial/vertical perspective) photos: Aerial Photographs of Mexico City Be sure to check out the 9th photo on the right; the first time I saw it I didn't think it was a real photo!
Yi-Tao - Mar 16, 2006 01:37:00 pm
Raymond- Those are some incredible shots. When you said to check out the 9th photo, are you refering to the low income houses? Anyways, for that photo, is there proof that it's real? Maybe it's because the houses are painted such bright colors, but the lighting and shadows don't seem realistic.
Ashley - Your article was really informative. When I took art history, they taught us that Renaissance painters used tools to help them with perspectives. So, I always thought that it was a given fact that they used tools.
Gwyu - Mar 16, 2006 08:57:58 pm
In class, there was some discussion about how having one perspective of a 3D scene on in 2D would not produce an image that necessarily matched the one constructed in the mind of a person looking at the same 3D scene, partly because the mind's image might be a composite of multiple views produced from the saccadic movements of our eyes. I was wondering that if this was the case, what is the size of the field of view used in creating these composites? For example, are the views foveal? It may be that knowing the size of the field of view would help determine the area that each perspective should have in a multi-perspective image.
Brien - Mar 19, 2006 10:39:27 pm
I especially like how the software in “Artistic Multiprojection Rendering” lets the viewer interactively arrange the perspectives in a scene to his/her preference. It would be interesting to do an experiment on multi perspectives like Blanz et al's – that is, instead of rotating an object, the user would select the perspective of a single object in a scene. I wonder what would be the most preferred perspective of that blue globe in Raphael's School of Athens. I kind of like it round.
