To label some sample images, first edit the IMAGES_ROOT
variable in
imagenetvid.cfg
to point to the root of your ImageNet Vid dataset.
FLASK_DEBUG=1 FLASK_APP=label.py FLASK_CONFIG=imagenetvid.cfg \
flask run -p 8000 -h 0.0.0.0
IGNORE BELOW FOR NOW; instructions below not updated from original video labeler used by @achald for a different project.
Once videos are labeled, you can filter the results by using the
filter_labels.py
script, like so:
python filter_labels.py \
--input-labels <path-to-labels> \
--must-have ${MUST_HAVE_LABELS} \
--must-not-have ${MUST_NOT_HAVE_LABELS} \
--output-labels <path-to-filtered-labels> \
--labels-list <path-to-labels-list>
In general, the you should use the following setup:
perfect - moving-unlabeled - MUST NOT HAVE static-labeled - MUST NOT HAVE no-moving-labeled - MUST NOT HAVE oversegmented-moving - MUST NOT HAVE static-part-labeled - MUST NOT HAVE interesting-video - look-again - MUST NOT HAVE challenging - always-moving -
In a sense, the 'perfect' label is kind of redundant; it just implies that moving-unlabeled, static-labeled, and no-moving-labeled are all false.