Closed avouros closed 7 years ago
Update: The user will have the option to visualize each trajectory and fix the arena center and the platform location accordingly.
Other addons (not implemented yet):
Partly completed in version 3.0.0 (commit: https://github.com/RodentDataAnalytics/mwm-ml-gen/commit/1ebfe0c31c7524a032e3ae0da3a223774123904b). Th user is now able to visualize the arena and move it until all the trajectories are inside.
Still missing the delete points and the interpolation options.
Calibration is performed before the acquisition stage. Afterwards it is locked (read only). If the user has the video files of the experiment available he may be able to redo the process and re-calibrate (takes a lot of time). If not then the only solution is to try to use the EthoVisions coordinates (world coordinates in the EthoVision versions XT 5+). Unfortunately this can be unreliable and needs more testing.
EPFL DATA: manually calibrated by us but center of the arena was indicated correct on the background image of EthoVision. The center of the platform didn't had a correct indication but it was easy to be detected )probably by some trial and error).
NENCKI DATA: The center of the arena in EthoVision 3.1 was problematic when it was detected from the background image. In EthoVision XT 10 and having specifying the center of the arena before the acquisition process everything was OK.
New EthoVision (XT 5+) EthoVision XT 5 was rewritten from scratch thus a lot of things has changed. Specifically, the calibration process is more straight forward but still the user has to specify (and take note) of the center of the arena and calibrate his data according to it. The center of the platform can be easily detected from EthoVision's background picture.