View images stored in photo_objects in the pickled numpy files (from the beetracker) - which are organised into a series of subfolders (system1, system2... etc, for example)
Adjust threshold scaling
Click on where the bee is (and add confidence note) & save this in a file
Remove previously labelled bees
View where these clicks have been
Web-based so it can be run remotely on a dataset
Move forward and backward through these photos (1 or 10 at a time -- 10 is important as often want to skip quickly forward/backward -- e.g. there might be 100s of 1000s of photos)
Move to a photo at a given time
Move to a photo of a given index number
Switch between tracking systems
Switch between cameras on tracking systems (the photos for different cameras will be mixed together in sub-folders)
Run the retrodetect algorithm on a series of photos
Display the output of the retrodetect algorithm
Display the output of alignment algorithm
Usage
I think a commandline tool is still good, as one can pass where it is happening, one can start it via a commandline on a remote server, etc. Ideally should work in multiple OSes.
edit: Out of date!
Example datafile:
29062023/cam1/photo_object_02G14695547_20230629_10:50:54.253289_flight2_0027.np
here we've saved the experiments from the 29th June, 2023, in a folder, within that there's 'cam1' (the first system) and the filename is of the form:
photo_object_CAMERAID_YYYYMMDD_HH:MM:SS.UUUUUU_LABEL_NNNN.np
where the LABEL is user set during fieldwork,
UUUUUU is microseconds
NNNN is the photo id number.
edit: See the new file structure for how we will name/organise files in future.
lionfish0
commented
3 months ago
Overarching functions
Usage
edit: Out of date! Example datafile:
29062023/cam1/photo_object_02G14695547_20230629_10:50:54.253289_flight2_0027.nphere we've saved the experiments from the 29th June, 2023, in a folder, within that there's 'cam1' (the first system) and the filename is of the form: photo_object_CAMERAID_YYYYMMDD_HH:MM:SS.UUUUUU_LABEL_NNNN.npedit: See the new file structure for how we will name/organise files in future.