search

darkpgmr / DarkLabel

Video/Image Labeling and Annotation Tool
316 stars 49 forks source link

MOT annotation format #5

Closed severous closed 2 years ago

severous commented 3 years ago

@darkpgmr How to convert to MOT's annotation format for multi-target tracking

darkpgmr commented 3 years ago

DarkLabel dose not support MOT's format. However, you can define your own format in darklabel.yml like that:

format_mot: 
  data_fmt: [fn, id, x1, y1, w, h]
  gt_file_ext: "txt"
  gt_merged: 1

And then, you can append "conf,-1,-1,-1" for each line after finishing your annotation.

Another way in case any noisy value for , , , are ok:

format_mot: 
  data_fmt: [fn, id, x1, y1, w, h, difficult, difficult, difficult, difficult]
  gt_file_ext: "txt"
  gt_merged: 1
darkpgmr commented 2 years ago

Now it support MOT format

QiyuLuo commented 2 years ago

@darkpgmr Thank your for your Open source tools. Since the MOT dataset is given in the form of pictures, when annotating it will appear save failed: format specifier 'fn' cannot be applied for images! How can I solve this problem？

Breezewrf commented 2 years ago

@darkpgmr Thank your for your Open source tools. Since the MOT dataset is given in the form of pictures, when annotating it will appear save failed: format specifier 'fn' cannot be applied for images! How can I solve this problem？

hello I got the same problem, have you solved it?

Breezewrf commented 2 years ago

@darkpgmr Thank your for your Open source tools. Since the MOT dataset is given in the form of pictures, when annotating it will appear save failed: format specifier 'fn' cannot be applied for images! How can I solve this problem？

hello I got the same problem, have you solved it?

got it it has to be video...