Because there are a lot of questions related to the image file format used by VirtualMirror here is a short answer…
At least a *.jpg of the frontal view of the eyeglasses is needed. This *.jpg should have a clean white background.
at least needed
Overlay, transparency, size, scaling, rotation, positioning will be computed by VirtualMirror itself.
Optionally VirtualMirror can use the Collada file format. This requires a web cam in any case. Both of the formats can be used together. VirtualMirror will try to enable the video mode if a Collada file (*.dae) was passed. Detailed information about Collada can be found at the following site: http://en.wikipedia.org/wiki/COLLADA
…written using Adobe Action Script 3 is amazing… I’m almost confused a little because of this.
We have a lot of installations around the whole world now. The software was translated to English, French, Italian, Spanish, Russian, Hebrew – and of course there is still a German version.
And we have extensions and code snipplets to embed this VTO into Magento, HikaShop, Shopify, OpenCart, Drupal, Joomla and sCat.
Even I’m about the html5 version since months I’m not very satisfied with the result, using Dart. That’s really confusing 
Since last week I was able to find a way to embed VirtualMirror into Shopify. I have “learned” Liquid for this… While the backend of VirtualMirror has to be run on an external server VirtualMirror itself can be called from cdn.shopify.com including the images to load.
However, I’ll prefer the “hard” way: A root server, running Linux + Magento.
…is here, of course this needs proof reading and many corrections.
It’s tricky really to have a right-to-left language using pure action script (AS3)
The kerning may be incorrect.
We have some more progress for the tablet html5 version even there are limitations, compared against as3.
…will be the next step. The development of the available AS3 version is continued.
Using HTML5 Virtual Mirror will have a reduced functionality compared against the AS3 version but at present this is an alternative way for tablet computers on client-side.
…was released as version 0.6.6 on Sunday, December 02, 2012.
The version is here.
The total width of the frame in millimeters can be passed as an additional parameter. Check out the different frame size (Medium Sized, Large Metal, and Kids Glasses).
Friday, December 07, 2012: The Bugfix Day
Apart from some small enhancements of the user interface two other longtime bugs are fixed:
- If the webcam was closed the default frame is loaded instead of the last visited one.
- If the web cam was enabled / disabled multiple times without a reload of Virtual Mirror the detection of face and pupils doesn’t work and the spinner is shown for ever.
The link on top was updated today, it may needed that you have to clean up your browser cache if you had gone to this page before.
…was released today. The tracking of the face / pupils / mouth on video is more exact. The robustness is a lot better.
A longtime bug “The Spinner doesn’t switch off sometimes also if detection was succesfully.” was fixed.
The API of VirtualMirrorDetector.swc itself is unchanged. Any previous version of Virtual Mirror will benefit from the power of the enhanced library.
As ever you can test the new version 0.6.5 here.
…was updated to version 0.6.5 in November 29, 2012.
Translation was done by Fabio Carli for the license of Ideolo.
Over this dark day – we got our first big snow here – I have written some code to embed Virtual Mirror into Hikashop, an extension for Joomla. While I have written our own e-commerce system in past using Perl and while I’m very familiar with Magento since the last year this was a new experience.
The best documentation of Hikashop is the php code of Hikashop
