RaFD paper accepted

Dear all,

We are proud to announce that the RaFD validation paper has been accepted by Cognition & Emotion. We are very excited about this. You can download the paper at Oliver Langner's personal website.

It will not be long now before we're ready to distribute RaFD to all of you. We have Bart van Delft, an excellent programmer who is studying Artificial Intelligence right here at the Radboud University Nijmegen, working on our website. We expect the website to be finished before christmas. We hope to be done with the final post-processing of all camera angles soon too.

We are looking back at a project that took three years of hard work for several BSI researchers, we should not let this go by silently.

Best,

Ron

RaFD near its final stage

Our last blog entry was a year old, but we have been working hard in the meanwhile to finalize the database. This is a general update post on the project.

1. We recently reserved the url that will host the final dataset, it will be www.rafd.nl. Currently, nothing is to be seen there, but a few students from the artificial intelligence department of the Radboud University are about to make a cool site for this.
2. The validation data have of course long been gathered and analyzed. The results look extremely good, with pretty high recognition rates. We hope to publish these data soon. A paper presenting RaFD and its validation is currently under review.
3. The legal issues have largely been resolved. A terms-of-usage contract has been set up and translated to english. If you're interested, you can see it here. In short, these are the most important details:
• RaFD will be usable for scientific research free of charge.
• Each researchers needs to register and download his or her own copy.
• RaFD can't be used for clinical treatment or other commercial purposes.
• Researchers need to be working at an accredited university.
4. The processing of most images is finished. As a small complication, we had to change our image alignment procedure (see earlier post) a bit. Specifically, we had to include a size fitting option for the pictures of the different camera angles. This was necessary, because the zoom of the cameras was not perfectly equal on all angles.

For all the researchers that have contacted us so far for using RaFD: We do not forget you! Once RaFD is in a usable state, which should be pretty soon now, we will inform you.

The RaFD-team