Preprocessing 3D Volumes for Tumor Segmentation using PyTorch and MONAI | Part 1/2

Поделиться
HTML-код
  • Опубликовано: 7 ноя 2024

Комментарии • 44

  • @pycad
    @pycad  Год назад

    Learn how to effectively manage and process DICOM files in Python with our comprehensive course, designed to equip you with the skills and knowledge you need to succeed.
    www.learn.pycad.co/course/dicom-simplified

  • @JimBob-lq1db
    @JimBob-lq1db Год назад

    Great Videos, helps me a lot to get in touch with pytorch and monai

  • @NikooMoradi
    @NikooMoradi 3 месяца назад

    It was really helpful, thank you very much.

  • @mardeenosman8979
    @mardeenosman8979 2 года назад

    Thank you I have been searching for this for a long time

  • @MuhammadMaab-os1gu
    @MuhammadMaab-os1gu 4 месяца назад

    can you explain.what labels exactly are

  • @futureme8887
    @futureme8887 2 года назад

    Thank you so much for this tutorial!

  • @mahdiebehjati9649
    @mahdiebehjati9649 5 месяцев назад

    Hi, Thanks for your video. Could you please explain more about four folders of Trainimages, Trainlables, Valimages and Vallables? I followed all the videos regarding the liver segmentation, but when got here I completely lost. you were talking about creating nifti files folder and suddenly jumped into four folders. what are difference between them? Is training folder the same files of niftifolder? what about validation data? where do they come from?

    • @pycad
      @pycad  5 месяцев назад

      You just need to do the split to the train and valid data. Most of the time, when we train a machine learning model we need to have the training and validation set separated from the test set for example.
      So you need to create the training and validation folders and name them the same that I did so that you don't need to change my training code and then it should work !

  • @nghethuatsong
    @nghethuatsong 2 года назад

    How can we have the dataset? Thank you for your nice video.

  • @favoritetroll9774
    @favoritetroll9774 Год назад

    Hey, can you tell me which organ you are working on in this video for segmentation of tutor

    • @favoritetroll9774
      @favoritetroll9774 Год назад

      tumor***\

    • @pycad
      @pycad  11 месяцев назад

      This was an example, you can do liver (for example).

  • @chillradio6338
    @chillradio6338 Год назад

    should i keep the data set in nifti format not dcom?

  • @truongtuankhue3156
    @truongtuankhue3156 5 месяцев назад

    Can you talk in detail about the 4 files TrainData, TrainLabels, ValData, ValLabels? For example how different it is

  • @cinthiakleiner820
    @cinthiakleiner820 3 года назад +1

    Hey. Nice video, thanks a lot for share. Can you guide where we can find a free dataset to try it ?

    • @pycad
      @pycad  3 года назад +2

      Unfortunately, I used data from the hospital in my project, so it is not available for sharing.
      There are, however, some publicly available data on liver, liver tumor, and lung tumor segmentation.
      Liver + tumor segmentation:
      www.kaggle.com/andrewmvd/liver-tumor-segmentation
      And these for lung and other tasks:
      medicaldecathlon.com/
      Because the liver dataset is for liver + liver tumor segmentation, there are three classes (background, liver, and tumor), so you must first remove the liver mask (if you don't need it) from the liver dataset.
      Good luck!

  • @kumarsai5607
    @kumarsai5607 Год назад

    Hi
    Thanks for sharing this tutorial
    Can you make tutorial on training a segmentation model using Monai

  • @med1343
    @med1343 3 года назад

    Bien venu de nouveau.

  • @fffgization
    @fffgization 3 года назад

    Hi. As far as I know for 3D images (x. y, z) its not height, width, number of slices. Most of the time, the x direction refers to going left/right on an image, y refers to front/back (or anterior/posterior), and the z direction refers to up/down (superior/inferior). but ur tutorials are great.

    • @pycad
      @pycad  3 года назад

      Hi Abir,
      First of all, I would like to thank you for your comment.
      I see what you are talking about (axial, sagittal, and coronal representations).
      You are right, but here I am working (or we can say I am displaying) the axial representation only which means it looks like we are having multiple images of 512x512 for that reason I am using the words width, height, and number of slices.
      Good luck :)

  • @hosammariee289
    @hosammariee289 2 года назад

    Dear brother, Thank you to this great toutrial
    my ques.
    I am trying to deploy my model but the code to prepare data or to preprocess the data
    train_transforms = Compose(
    [
    LoadImaged(keys=['image']),
    AddChanneld(keys=['image']),
    Spacingd(keys=['image'], pixdim= (1.5,1.5,2)),
    ScaleIntensityRanged(keys=['image'], a_min=-900, a_max=1000,b_min=0.0,b_max=1.0, clip=True),
    CropForegroundd(keys=['image'], source_key='image'),
    Resized(keys=['image'], spatial_size=[128,128,90]),
    ToTensord(keys=['image', 'label'])
    ]
    )
    work on a collection of data I need to take an image from the request and per-process it
    can you help me

  • @hayetberrah8579
    @hayetberrah8579 3 года назад

    Bon contunuation👍👍👏👏👏

  • @Nature_Traveller-PR
    @Nature_Traveller-PR Год назад

    Brother l need medical image data set can you please help in this regard

    • @pycad
      @pycad  Год назад

      What kind of dataset do you need?

    • @Nature_Traveller-PR
      @Nature_Traveller-PR Год назад

      Radiology images along with reports like chest x- ray, kidny abdomen etc for my research point of view

  • @ivotje12
    @ivotje12 3 года назад

    Hi,
    thanks for your tutorial! great!
    first question:
    what do you mean by train labels? what is your label data?
    second:
    are there coming more tutorials on this subject? I love it
    thanks in advance!
    Ivo

    • @pycad
      @pycad  3 года назад +1

      Hi Ivo
      For the first question:
      → I am doing tumor segmentation in the whole body, so my labels are binary masks (images with 0 and 1).
      For the second question:
      → Yes there is the second part of this tutorial, which makes it clear for you in this link
      ruclips.net/video/hqgZuatm8eE/видео.html
      → In the next video I will talk about how to do data augmentation in these 3D volumes.
      → And I have already blogs and videos about these kind of stuff (python and computer vision)
      - You can find the videos in my channel.
      - Or the blogs in my website, here: pycad.co/
      Good luck and happy learning.

  • @akainu3668
    @akainu3668 3 года назад

    thanks a lot ,can you please make more of these for monai

    • @pycad
      @pycad  3 года назад

      Glad that you liked it.
      For now there is the second part of this videos that you can find here:
      ruclips.net/video/hqgZuatm8eE/видео.html
      And there is my blog post about doing 3D volumes augmentation using Monai (the video will be here soon).
      pycad.co/3d-volumes-augmentation-for-tumor-segmentation/

    • @akainu3668
      @akainu3668 3 года назад

      @@pycad thanks a lot, is there a way to get your advice I need to learn to work on monai as my project requires it

    • @pycad
      @pycad  3 года назад +1

      @@akainu3668 Yes of course.
      You can send me an email or a message in my linkedIn
      I will be glad to help 😄

  • @anima_kujur
    @anima_kujur 3 года назад

    Hi! Nice tutorial.
    Could you please help me with 3D tooth segmentation. I have 3D CT images of teeth and it is in the .STL file format. I don't know how to preprocess it. Couldn't find much help from Google search.
    If I can get any help from you, I'll be very grateful.

    • @pycad
      @pycad  3 года назад

      Hi,
      I am sorry to tell you that I do not have any idea about this kind of file, unfortunately.
      Good luck

    • @anima_kujur
      @anima_kujur 3 года назад

      @@pycad thanks for quick response. No problems

  • @nayrapumar1324
    @nayrapumar1324 3 года назад

    Is there any way to download the .nii files used in this example to try to replicate it locally?

    • @pycad
      @pycad  3 года назад +1

      Unfortunately, I used data from the hospital in my project, so it is not available for sharing.
      There are, however, some publicly available data on liver, liver tumor, and lung tumor segmentation.
      Liver + tumor segmentation:
      www.kaggle.com/andrewmvd/liver-tumor-segmentation
      And these for lung and other tasks:
      medicaldecathlon.com/
      Because the liver dataset is for liver + liver tumor segmentation, there are three classes (background, liver, and tumor), so you must first remove the liver mask (if you don't need it) from the liver dataset.
      Good luck!

  • @abhinavkumarmishra5272
    @abhinavkumarmishra5272 Год назад

    Hello sir, while doing the code of pre-processing part, i have encountered an issue "FutureWarning: : Class `AddChannel` has been deprecated since version 0.8. please use MetaTensor data type and monai.transforms.EnsureChannelFirst instead.
    warn_deprecated(obj, msg, warning_category)"
    After this i have replaced the AddChanneld to EnsureChannelFirstd and also imported the library of MetaTensor. while executing above scripts, my kernel was dead and it doesn't reflecting any thing other than this. Even i dont know where i have to use the metatensor library. Kindly help me resolving this issue sir.

  • @mokhtariali5097
    @mokhtariali5097 3 года назад

    Nice

  • @YouCef-xg5ps
    @YouCef-xg5ps 3 года назад

    خويا ربي يرحمك عمري 17 حبيت نسقسيك دك انا عندي هادا العام نجوز الباك ادا لازم نسنا حتى مور مانجوز الباك باش نقدر نبدى لي ديمارش ولا نقدر نبدا دكا مام بلا باك ؟

    • @pycad
      @pycad  3 года назад +2

      Salem khoya,
      Kayen li yadbaw ydiro les démarches fel 3am ta3 l bac bach yjiw ya9raw L1 hna direct mais s3iba bach ya9blok sama dd préférence hta mor l bac

    • @YouCef-xg5ps
      @YouCef-xg5ps 3 года назад

      @@pycad merci beaucoup khouya

    • @YouCef-xg5ps
      @YouCef-xg5ps 3 года назад

      @@pycad khouya la3zize ana habite za3ma nsayi hada la3am bache nadi experience .
      تعيش خويا السآل هو : فالفيديو التاعك خويا قلت بلي لازملي الباك باش ندير ليديمارش كيفاش نبدى دوكا ندير ليديمارش بلا باك ؟ 🙏🙏🙏🙏🙏

    • @pycad
      @pycad  3 года назад +1

      @@YouCef-xg5ps les démarches ta3 wahd L1 w les démarches ta3 lokhrin machi kifkif.
      Ntoma 3andkom un test de français spécial asmah DAP w procédure un peu différente 3la les autres. Ana ma3andich fikra complète 3liha mais ta9dar tchof m3a ccf.

    • @YouCef-xg5ps
      @YouCef-xg5ps 3 года назад

      @@pycad merci bcp merci bcp khouya hakda wlate 3andi nadra merci bcp merci bcp merci bcp