GitLab

Code further profiled, and removed a couple unnecessary operations.

Added information about possible move to OpenGL for the camera widgets.

Fixed bug for eye processor pre-processing downscaling.

Since OpenCV seems able to handle other containers now, we are moving away
from .avi (and it's small size limitation).

I experimented with 3 codecs for static and dynamic scenes and had mixed
results when comparing MPEG4 and H264; based on the Surface results (least
powerfull test device), I'm going with MPEG4 to begin with.
This choice is also better from the usability side since the user no
longer has to download the Cisco dlls.

MJPEG is faster, but the video size becomes huge too fast. In the future,
we make this parametrizable for the user similar to Pupil Labs.

On an additional note, the generated .mp4 videos are playable with Pupil Player
v0913.

Moving EyeRecToo to a public repository.
The commit history was removed as it may contain sensitive information.

Swirski and Starburst pupil detection algorithms are not included by default and
will be made available on separate repositories when the new pupil detection
plugin structure is included.

For internal use:

This is a reorganization of the build structure from:

Repository: EyeRecToo-internal
Branch: migration
Commit: a2da2fbbcb7521b538c81b9edfc855404702b35b