dbolkensteyn/vatic.js

vatic.js A pure Javascript video annotation tool

vatic.js's Issues

如何去掉解帧的那部分，放入自己的照片

No license for code

There is no license for this code, and that makes it hard for me to modify, contribute or reuse the code for my own purposes. If you are happy to share, consider maybe an MIT license? Github maintains a site describing the MIT license and other licenses here: https://choosealicense.com/. Thanks for making the cool code available!

Format of ZIP file for extractFramesFromZip?

What is the appropriate format for the ZIP file when using extractFramesFromZip?

Can I just use it for offline?

when I download this JS tool for video annotation, many problems occured. we cannot use this tool offline,right?

Loading custom extracted frames

I have extracted with a script the frames from a video, the images are of .jpg Format and when I upload them I get the following message: Extraction completed. 0 Frames captured, the image names are index.jpg (i.e 0.jpg, 1.jpg and 2.jpg)

Can't download frames zip file

[Violation] 'click' handler took 483ms

how to use polygon

Implementing other trackers

Is there any roadmap to implement other tracker like KCF?

What is the xml export file type? Any way to convert to PASCAl VOC?

I assumed the xml export would be a vatic xml format but the only information i can find on that format shows a different xml structure than what vatic.js exports, what is the official xml format of vatic.js export? (The export looks closer to VOC but not separated into files per image/frame: https://eavise.gitlab.io/brambox/api/boxes-annotations.html#pascal-voc)

Also, my goal is to convert to pascal voc, do you have any info that could help or point me in the right direction?

BTW @dbolkensteyn the optical flow is amazing and in general the speed that i can annotate with this tool is super fast- I haven't found another tool that is as fast!

Control sampling rate

I would like to use this tool to extract some frames from image to train deep learning system.
The issue i have is too many similar frames generated.
Would it be possible to set constant sampling rate for video?
I.e show and annotate only two frames per second?

Working offline?

I am able to run add bounding boxes and then have optical flow track them through the video when I use https://dbolkensteyn.github.io/vatic.js/. Although, when I use a local version downloaded on my computer, I can only read the frames and add bounding boxes, but the video doesn't play, neither does the tracking work. The video only plays when I remove/hide all the bounding boxes. I need the tool to work offline and locally for the data that I will be using it for.

I suspect I missed something that's needed to set it up to work offline. Any support would be highly appreciated.

frame rate hard coded ?

File or resolution limit?

This tool is very useful, thank you!

However, I'm trying to upload a 1080p video for annotation that is 1 minute long (38.5 MB) and the frame extraction never starts. I've successfully annotated 720p videos that are larger in file size. Is there a limit to the video resolution?

This looks great!

Hi I just tried this tool and its amazing!
I have another project https://www.eraseimage.com/ which does superpixel segmentation in JS.
One of the big advantage of this project is that it implement a nice annotation UI using Fabric.js.
You can adopt this for Vatic https://github.com/AKSHAYUBHAT/ImageSegmentation