I've created a show-off app for my recent implementation of face detection in flash. If you've got a webcam, you can play with it now. It's also got "Load from URL" functionality, but you've got to find a server that has a permissive crossdomain.xml file. Supposedly flickr and yahoo have appropriate permissions, but I've not tried them.
If you're running firefox, the "export to page" option should insert the altered image into the html so you can right-click-save it. IE doesn't seem to work. I may integrate export to ImageShack in the near future.
Anyone interested in helping with optimization of the face detection library? (I'm talking to you: rachil0, MyFriendIsATaco, 691175002, and others)
So that this thread isn't all about me, please discuss your own projects which show off cool stuff with really dumb demos.
You should start the application out with an image already loaded in for us to play with. :P You're probably logging all the webcam footage. I know I sure as hell would be for something like this!
And why is this directed towards me? Haha! Think I can improve on this? I need to see something working first.
In my local testing, loading up images worked fine. I didn't discover that nocrossdomain.com didn't work for images until after I'd uploaded it. I'll add some sample images to play with tomorrow.
I'm not logging any webcam footage. I'd need to set up red5 for that, and that's way too much of a headache right now
And I just listed a few of the more technically able coders I thought might read the thread. I know that there's plenty of room for optimization, and I'm planning to open up the source soon. I wanted to get an idea of whether anyone would want to work on it.
Also, how are you doing the export to page script? I don't see why it would only work in Firefox.
[edit]
Oops, I see now that you are just dumping raw bytes into an <img> element. I guess Firefox is the only one to support that? I thought it was a standard for all browsers.
[/edit]
I'd be interested in checking out whatever algorithms you're running...I may see some tweaks that could be made to improve performance. Of course I'm not well establlished with this forum quite yet so no worries if you don't want me to >_>'
Anyone interested in helping with optimization of the face detection library? (I'm talking to you: rachil0, MyFriendIsATaco, 691175002, and others)
Swamped at the moment, but really flattered to be mentioned! Haven't checked it out yet, will view it tonight from a ... nonwork location that is more suitable for lingered supermodels. Can't wait to see it.