Below you'll find data from the scanner project as progress is made. Order is reverse chronological with the most recent stuff on top.

software you can use to view this data:

• ivview, an OpenInventor-based viewer. Supports points, IV files, and vrml1
• PointCloudsShowCenter.java - my simple Java3D multiple point cloud viewer which reads raw ascii points files with either a single header line indicating the number of points in the file, or two lines containing xres and yres respectively (actually order doesn't matter). Also this simple script to launch it 'cause invoking java manually is tedious.
• Alias Wavefront: alias, evalview. find 'em wherever fine SGI's are sold.
• FreeWRL - a shitty free VRLM2-only viewer. May be useful as plugin infrastructure for other apps, but used alone it seems to suck.
• gnuplot? sure, why not.
• Geomview - reads OOGL files, comes with a vrml1 converter
• Images listed here are all jpegs for size, but in actual operation are processed as pgm or ppm. Most images here are rotated from the portrait orientation

Hand-held Rotary Cheese Grater (scanned manually, 4/1/01)

72-frame manual scan. This came out surprisingly well. Unfortunately there was some weird artifact in the vrml conversion which I can't figure out (which is not evident in the screenshot, but you'll see it in the vrml model).

• a color image
• screenshot of the pts model - note the handle on the left.
• screenshot of the vrml model
• raw points
• points after editing in evalview
• vrml1

Water Pistol 4 (scanned manually, 4/1/01)

72-frame manual scan. Another go a this one with more success.

• a color image
• screenshot of the pts model
• screenshot of the oogl model
• raw points
• points after editing in evalview
• oogl
• vrml1

Water Pistol 2 (scanned manually, 3/18/01)

72-frame manual scan. I offset the water pistol from the axis of rotation by about 3cm - the max possible before the end of it crashed into the arm of the turntable. Distortion on the turntable (had to replace the camera batteries which requires a tilt and I didn't recalibrate) and the pistol itself points to errors in my camera calibration and hence data reconstruction. Offsetting the object did wonders for lateral coverage, however.

• a color image
• points
• vrml1
• sample scan photo - butt of the pistol (see first image)
• sample scan photo after processing

Water Pistol 1 (scanned manually, 3/16/01)

This 72-frame manual scan was greatly helped debug the data. Notice how the center of the pistol is very densely covered as it is on the axis of rotation, at which the laser is trained.

• a color image
• points
• vrml2
• sample scan photo - the top of the pistol (see first image)

Leather Wallet (scanned manually, 3/16/01)

This was the first manual scan which let me work out a bunch of bugs. The turntable is extremely clear and the wallet, opened slightly in order to stand, can be seen perched on top. There are some artifacts and a couple scanlines of the end-view.

• a color image
• points screenshot - notice the timing flange extending from the left edge of the turntable
• points
• sample scan photo - note how dim the line on the wallet is

Susan Ivannova (scanned automatically, 3/9/01, processed more recently)

This automatic 204-image scan was performed at roughly 27rpm and 10fps. You can see how inconsistent the turntable and/or the time between frames is. Motion-blur is also a horrendous problem.

• a color image
• points

Pot (scanned automatically, 3/9/01, processed more recently)

This automatic 244-image scan was performed at roughly 27rpm and 12fps. The pot is a wheel-thrown (by me) pot which was placed upside-down on the axis of rotation to serve as a test by scanning a symmetrical object. The banding in the data is due to the multiple rotations giving increasingly more coverage.

• a color image
• points - had to generate this w/ a sub-sampled point set
• vrml2