search

cvondrick / vatic

Efficiently Scaling Up Video Annotation with Crowdsourced Marketplaces. IJCV 2012
http://mit.edu/vondrick/vatic/
MIT License
608 stars 255 forks source link

Precise Annotated data #101

Open YangBain opened 6 years ago

YangBain commented 6 years ago

@hepek @marcovzla @Nikhil-Kasukurthi @cvondrick Recently, I use vatic to annotate video data. But when I use F and D for forward and backward, the results are not very good. So, in this case, who knows how to get the precisely annotated data. And if I use V and C directly, I can not save the results. So could you give me some suggestions for how to get the precisely annotated data. Thank you very much.

gosowski commented 6 years ago

@YangBain I have exact the same problem. Did you figure out how to solve it ? Edit: I solved the problem. The solution is to change track.js file from vatic public directory to track.js file from @onauparc pull request #20. What is worth to notice is to refresh browser (shift+f5), after changing track.js files.

YangBain commented 6 years ago

@gosowski I am sorry I do not solve this problem. Finally, I use another image annotation tool to do annotation work. Vatic always output many wrong bounding boxes for my case.

1329844541 commented 6 years ago

another image annotation tool ？What tool?