fhennig / face-of-the-crowd Goto Github PK

Either there needs to be a fullscreen portrait view with a smaller image of what the webcam currently shows, or we could already overlay the face of the person standing in front of the picture and then have a button to allow them to save their image to the collection of images.

That would be even more "magical".

The artwork still needs a title and a description/background and how it relates to the topic "artsci, illustrating research."

There should be a "cooldown" afterwards so the same person isn't in the picture multiple times.

It should work given the the frame of the camera is already quite narrow.

At the press of a key, the current image of the webcam should be saved to disk.

Face poses can be extracted as in here: https://www.learnopencv.com/head-pose-estimation-using-opencv-and-dlib/

When an image is saved, before it is saved, it should be scaled and maybe rotated and/or "prespective distorted" to make every face more similar in alignment.

Something like this: https://www.pyimagesearch.com/2017/05/22/face-alignment-with-opencv-and-python/

In the end the saved images should be able to just align smoothly.

Instead of pressing a button, users could make a hand gesture (thumbs up) to interact with the device.

stuff to install on raspberryPi: https://stackoverflow.com/questions/53347759/importerror-libcblas-so-3-cannot-open-shared-object-file-no-such-file-or-dire

https://stackoverflow.com/questions/22654770/creating-vignette-filter-in-opencv/22843077#22843077

faces are saved into a directory.

https://pypi.org/project/imageio/

Use the mask parameter in the add function.

Frame f, m;

for g in otherFrames:
    f = f + m * g
    m = m * mask(g)