A quick but comprehensive guide to LabelMe - an image/video annotation tool for deep learning

Поділитися
Вставка
  • Опубліковано 24 січ 2025

КОМЕНТАРІ • 32

  • @JuanaMike
    @JuanaMike 2 роки тому +1

    I was looking for this all over the internet. Thank you

  • @priyankaarora946
    @priyankaarora946 3 роки тому +3

    very easy and informative video. please explain about advanced features of labelme that you mentioned

  • @juanmontenegro7239
    @juanmontenegro7239 3 роки тому +3

    Really cool video, thanks!
    A nice idea would be on how to create a custom instance segmentation and/or object detection small dataset and train it

    • @ovision01
      @ovision01  3 роки тому

      Thank you for the comment and suggestion. I’ll soon upload a video on how to train Yolov5 (object detection model) on a custom dataset. Please, stay tuned and thank you for your support once again.

  • @pavithraprashanth2313
    @pavithraprashanth2313 Місяць тому

    Thank you very much!!You are amazing!

  • @saborihusain2025
    @saborihusain2025 2 роки тому +1

    Very well explained. Thank you so much. Is there a video coming out about the advanced features of LabelMe?

  • @seker378
    @seker378 2 роки тому +1

    Thanks mate. This is a very helpful video for me.

  • @sanghamitrajohri2596
    @sanghamitrajohri2596 2 роки тому +1

    How to change the colour of the images that are getting annotated, say I needed the sofa in orange box instead of grey

  • @pwts2004
    @pwts2004 10 місяців тому

    Could you please share how to train the model with created dataset? it will be very helpful

  • @shirshahzarenkhi1262
    @shirshahzarenkhi1262 2 роки тому +1

    well done brother .

  • @anopta
    @anopta 3 роки тому +1

    Wow. That’s very informative

  • @jdiazram
    @jdiazram 2 роки тому +1

    Hi, nice tutorial, I have a question: the 4 numbers of points from json are x1,y1 and x2, y2 from de bounding box? Thx

  • @vijaypatil8854
    @vijaypatil8854 2 роки тому +1

    nice tutorial I have question if there are multiple similar items in an image then how to label them all (say multiple seeds or pods)?

    • @ovision01
      @ovision01  2 роки тому

      I don’t get your question well. Do you mean similar objects or same objects? Let’s say there are two cats in an image. Both should be labeled as “cat”. Hope it helps

  • @awadelrahman
    @awadelrahman 6 місяців тому

    Thanks! Very much appreciated :) BTW, is it still the best free tool for this purpose? Any other tools you might suggest?

  • @user-nc6vb1gf4u
    @user-nc6vb1gf4u 7 місяців тому

    *HELP*
    Hi, thanks for this video, I am annotating lots of photos to feed software. I would like to modify a label in the list. how to proceed without doing it on each photo one by one? thanks in advance

  • @ninadchobe6044
    @ninadchobe6044 11 місяців тому

    Can you suggest what to do if I want to label 1000s of images for the same label type. I have to click and go through the pop up millions of times. Please, if you could help I would greatly appreciate it

  • @geospatial90
    @geospatial90 11 місяців тому

    Sir how can I convert json to text or png to text for Yolo8 model? Can you please help this...

  • @SalaarArifKhan
    @SalaarArifKhan 2 роки тому +1

    Is there a way to resize the images after annotating in LabelMe ??

    • @ovision01
      @ovision01  2 роки тому

      Could you elaborate your question a little bit more?

    • @SalaarArifKhan
      @SalaarArifKhan 2 роки тому

      @@ovision01 once the images are annotated with polygons marking regions of interest, is there a way to resize the images ALONG with the resizing of the regions of interest?

    • @MS-bs4cg
      @MS-bs4cg 2 роки тому +1

      @@SalaarArifKhan I’m afraid there’s no such feature available in Labelme. However, you can easily do that using Python with openCV. All you need to do is to resize the image using a ratio and apply the same ratio on your labeled area enlarge it. Hope it helps

    • @SalaarArifKhan
      @SalaarArifKhan 2 роки тому

      @@MS-bs4cg i tried exactly this. But the newly resized images do not show up in their new sizes in labelme. Only the new annotations show up correctly. So correct annotation on old image size is visible on labelme which is a bit frustrating.

    • @SalaarArifKhan
      @SalaarArifKhan 2 роки тому

      @@MS-bs4cg even with new image dimensions, labelme is reading the old hash of the image and does not show up image with new size

  • @aycatekin8521
    @aycatekin8521 2 місяці тому

    how do we write a code to convert a json file into an image WİNDOWS

  • @theodor-alexandrutufa7545
    @theodor-alexandrutufa7545 2 роки тому +1

    Hi, currently the latest release is 5.0.2 but I'd like to use the 4.5.12 - could you please guide on how to install that version?
    I have tried pip install labelme v4.5.12 - Nothing
    pip install labelme 4.5.12 - Nothing..

    • @ovision01
      @ovision01  2 роки тому

      Use pip install labelme=4.5.12. Hope it solves your problem

  • @neelotpaltripathi841
    @neelotpaltripathi841 2 роки тому

    bro i labelled the data but when same objects are their in different images it gives different colours to it instead of same colour

  • @inquisitiverakib5844
    @inquisitiverakib5844 2 роки тому +1

    thanks a lot