search

Cartucho / OpenLabeling

Label images and video for Computer Vision applications
Apache License 2.0
926 stars 266 forks source link

load bouding boxes from previous images. #62

Open YeetMasterYeet opened 5 years ago

YeetMasterYeet commented 5 years ago

hi, i have a small suggestions to enhance open labelling, for video usage the frame boxes are basically the same, i suggest that we should add a feature that allows to load the bouding boxes from the previous frame and add a button to move the box around so that it become muc hquicker the treat data from videos. cordially.

fogx commented 5 years ago

you can use a tracker to track multiple frames rapidly. Once the tracker diverges from the target just delete that frame and all following get deleted as well.

zhangqinghao0811 commented 4 years ago

Yes, I can't agree with you any more. It's very useful for tracking a motionless object, as DaSiamRPN tracker is not stable sometimes.

zhangqinghao0811 commented 4 years ago

you can use a tracker to track multiple frames rapidly. Once the tracker diverges from the target just delete that frame and all following get deleted as well.

Look this example, the turck is motionless, but the tracker's prediction is unstable.

fogx commented 4 years ago

the example is restricted, i can't see it. You might have to try some of the other trackers then. I prefer the CSRT Tracker

zhangqinghao0811 commented 4 years ago

the example is restricted, i can't see it. You might have to try some of the other trackers then. I prefer the CSRT Tracker

Sorry, try this link, thanks for your advise, I'll try.