Meet Jill. Click on her image to activate a miniPicView with a 3-D PicTree of Jill (from a color laser scanner). (If you've configured your WWW browser correctly, the viewer will automatically start.) (Note: The images used to bring up a PicTree viewer may exhibit color problems because sometimes a fixed color table is used for all pictures in the document. This will not be a problem for the viewer because it loads a custom color table.)
To make things exciting, we set Jill to come up rotating. You should be warned that, because the object only represents Jill's face, you may see some very strange images (when you're viewing both inside and outside).
You can still use the sliders and mouse buttons while it's moving. Use the left slider to make Jill smaller (down) to increase performance (regardless of the window size). This is an inherent characteristic of the hierarchical PicTree data structure. Less information needs to be processed when generating a lower resolution image. Also, if the object is small, making the window smaller will improve performance (because the window system has less pixels to move around).
You may want to turn tumble off by pulling down the Tumble Menu and selecting (surprise!) Stop. You can reset the scale and viewpoint by selecting Reset from the same menu.
It should be noted that the initial viewpoint, scale factor, window size, tumble mode (on/off) and other parameters are contained in RPL instructions in the Jill PicTree. They could have been set to any values.
The original Jill data set has 24-bit color values. Because most systems that will be used to view this demo use 8-bit displays, Jill was dithered down to 256 colors. When you scale it up, you can clearly see the limits of this technique from some of the strange colors. A full-color Jill Pictree is also available.
The highest resolution in this PicTree is one part in 256 (the lower resolutions are also represented; it's hierarchical!). It occupies 66.5K bytes of storage. We don't think that's too bad when you consider that an uncompressed 256 by 256 image (8 bits per pixel as with Jill) would require 64K bytes. A lower resolution Jill PicTrees would require less memory (with lower image quality).
If you select the Jill image below, a viewer with a higher resolution (one part in 512) Jill will pop up.(If it doesn't come up, we probably sent this to you on floppy disks and were too cheap to add another disk.) At lower scale factors, the image will be the same (it's using the same information). At larger sizes, however, a higher quality image is generated. This requires a PicTree about four times larger (is's 203.6K bytes). Try scaling the image up (left slider up). Note that you can eventually see the actual resolution of the scanner (a 512 by 512 by 512 cubical array of voxels). You may want to put the two side-by-side for a quality comparison (is the extra memory worth it?). Actually, the Jill PicTrees could have been scaled to any resolution to precisely trade off memory size and image qualtiy when they were made. These are just two examples.
The picture below is of Dr. J. Clodd Nurdsworth, one of the original pioneers in the selective breeding of PicTrees (from octrees and, before that, oak trees). Click on the image to bring him up in 3-D.
Note that Dr. Clodd is not just a pretty face (like Jill) but part of a shirt and lots of hair also. Note: If the viewpoint controls are cleverly manipulated to view up the neck from below, one can verify that there is absolutely nothing inside his head (something that has been suspected, but not proven, for many years).