Week5 : 3D Scanning and Printing

Using some of the many available open source codes for photogrametry I was able to test this technigue on several images and also a drone flyover video.

The most stable to use is VisualSFM (Shape From Motion) by Changchang Wu. It also bunddles other key algorithms to produce dense reconstruction.

The following is the result of 46 images I found online for a historical theater ruins in Syria in the historical site of BUSRA ( at the time of Testing we were involved in an effort with Arab regional center for world heritage ARCWH to establish a repository of images and a data and attempt to reconstuct some of the site in 3D as a conservation effort).

I wanted to test this fab-modules function and needed to find a suitable GIF file. I searhed for animations

but obviously this was not the best example

gif_stl sphere.gif sphere.stl

I have used dicom ( a standard file extension for medical imaging files) previously and thought a ct scan data would make a nice gif animation. There are many radiological datasets online, but this time I have used the first link I got in the goolge search.

http://www.osirix-viewer.com/datasets/ Osirix is a open source dicom image processing software for radiological data visualization and processing.

Other more usfeful link is the following https://mri.radiology.uiowa.edu/visible_human_datasets.html

I aim to use the data set to reconstruct a 3D mesh. I wanted to try using the fabmodule gif_stl command. obviously the first attempt using the above sphere gif was not the best example. That's why I tried a set of CT scaned images.

Using a simple python script I read the dicom images converting the saved jpeg to a gif animation, since there is no module or a straight forward way to do so in python I used a much easier command line conversion method using ImageMagick library.

convert -delay 100 -loop 0 image*.jpg animation.gif

Then using the fabmodules also through comand line I got the stl file

and using a simple threshold segmentation in python I get the bone only

dicom-jpeg

dicom-jpeg (segmented bone)