(Ln(x))³

The everyday blog of Richard Bartle.

RSS feeds: v0.91; v1.0 (RDF); v2.0; Atom.

Previous entry. Next entry.

11:28am on Saturday, 4th March, 2006:

3D images

Comment

I was talking to my elder daughter the other day about 3D vision. Recalling my old AI lectures, most of it (60% or so) comes from having two eyes (through stereoscopic vision and, a little, through convergence — if I need to cross my eyes more to look at the ball than the bat, the ball is closer than the bat). However, you can also figure out a 3D scene if you use just one eye:

• Occlusion — I can't see all the man because there's a car in the way, therefore the car is in front of the man.
• Focus — I need to tighten my focus to see the pencil more than when I see the mug, therefore the pencil is closer than the mug.
• Knowledge — Houses are much bigger than dogs but these look to be the same size, therefore the dog is closer than the house.
• Motion parallax — When I drive past the pylons they move faster than the trees, therefore the pylons are closer than the trees.

There are a few others, too (eg. the colour of things far away seems less intense than that close up), but these are the main ways of knowing where things are relative to one another in space from looking at them.

Now for years, I've taken 3D photographs. Here's an example of Stonehenge in 3D. To view it, you have to cross your eyes to make a single image, then bring this image into focus. The full technique is explained here.

Talking to my daughter, though, it occured to me that I could also show my pictures in 3D using movement. After some experiment, I came up with this. It's not bad, although it's a little irritating after a few seconds. You get the 3D effect, though.

I'm not going to convert all my 3D pictures to this format, however, because it's quite brittle when compared to the cross-your-eyes method. Changes in colour, focus, angle of rotation, framing and proximity to the subject all show up a great deal more than they do with stereoscopic vision (which is very robust).

Still, it's an interesting exercise.

Referenced by 3D from 2D.

Latest entries.

Archived entries.

About this blog.

Copyright © 2006 Richard Bartle (richard@mud.co.uk).