Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
    Leon Wolf
    @fogx
    hi everyone, i just checked out your project. It looks very pretty. thumbs up for that design implementation!
    I was sad to not find an import function though. Is there no way to import previous annotations into datasets?
    my favorite labeling tool: openLabeling has a very ugly UI, but amazing features (openCV trackers, even a DL feature). Also the order structure with input/output makes alot of sense to me. Maybe you could convert some of that into your own project.
    cheers
    Piotr Skalski
    @SkalskiP
    Hi @fogx ! It's not here yet. But for a long time now, we have had requests to do so. On GH, an issue looking for this topic has the most people participating in the discussion.
    As far as OpenLabeling is concerned, it is worth noting that this repository was created at the beginning of 2018. I wonder where they were after 3 months ;)
    We also have ambitious feats in our plans :), but there is not enough time and maybe another person to help.
    Leon Wolf
    @fogx
    yes i definitely agree. But it gets some things quite right, so it might be nice to take a look :)
    Piotr Skalski
    @SkalskiP
    You're definitely right! There are a few cool projects that are good to get inspired from.
    I don't know if all these projects are possible at JS. But I will certainly look around and try.
    sreevardhanreddi
    @sreevardhanreddi
    hi everyone, i just checked out your project. It looks great, I need some help. I am trying to customize it, I just want the editor and labelling part.
    AShabirG
    @AShabirG
    Hello, I was using the MakeSense web page to annotate some images and it was working but when I come back the next day with a larger sample of photos it stopped detecting the required objects. Is there anything I can do as I need a few hundred to train a network. Thank you
    Piotr Skalski
    @SkalskiP
    Hi @AShabirG could you describe your problem in a little more detail? What stopped working?
    AShabirG
    @AShabirG
    Sure, so the AI model was able to detect pretty accurately the position of cones in the images but it is now not able to detect anything in the images. I inputted the original images again and it still yielded no output.
    Piotr Skalski
    @SkalskiP
    we talk about the models that are available inside MakeSense?
    could you send me some example image?
    AShabirG
    @AShabirG
    image.png
    When using the coco ssd ai this is the output for numerous examples whereas previously I would get outputs similar to the next image
    This image I'm about to post is one I annotated myself but I did not screenshot examples where the annotation worked correctly but I do have csv files for them.
    image.png
    AShabirG
    @AShabirG
    Sorry for the lack of detail, I am a novice when it comes to this
    Piotr Skalski
    @SkalskiP
    Could you also give me raw image? The one that you upload to make sense? Not a screenshot? :)
    AShabirG
    @AShabirG
    frames_2580.jpg
    If you require any more just ask, I have a few thousand