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
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?
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 !
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!
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.
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 :)
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
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
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.
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/
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.
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!
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.
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 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.
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
Great Videos, helps me a lot to get in touch with pytorch and monai
It was really helpful, thank you very much.
Thank you I have been searching for this for a long time
can you explain.what labels exactly are
Thank you so much for this tutorial!
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?
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 !
How can we have the dataset? Thank you for your nice video.
Hey, can you tell me which organ you are working on in this video for segmentation of tutor
tumor***\
This was an example, you can do liver (for example).
should i keep the data set in nifti format not dcom?
Can you talk in detail about the 4 files TrainData, TrainLabels, ValData, ValLabels? For example how different it is
Hey. Nice video, thanks a lot for share. Can you guide where we can find a free dataset to try it ?
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!
Hi
Thanks for sharing this tutorial
Can you make tutorial on training a segmentation model using Monai
Bien venu de nouveau.
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.
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 :)
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
Bon contunuation👍👍👏👏👏
Brother l need medical image data set can you please help in this regard
What kind of dataset do you need?
Radiology images along with reports like chest x- ray, kidny abdomen etc for my research point of view
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
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.
thanks a lot ,can you please make more of these for monai
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/
@@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
@@akainu3668 Yes of course.
You can send me an email or a message in my linkedIn
I will be glad to help 😄
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.
Hi,
I am sorry to tell you that I do not have any idea about this kind of file, unfortunately.
Good luck
@@pycad thanks for quick response. No problems
Is there any way to download the .nii files used in this example to try to replicate it locally?
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!
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.
Nice
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 ؟
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
@@pycad merci beaucoup khouya
@@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 ؟ 🙏🙏🙏🙏🙏
@@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.
@@pycad merci bcp merci bcp khouya hakda wlate 3andi nadra merci bcp merci bcp merci bcp