{ "id": 370045, "title": "There are two types of Dicom images in dataset.", "url": "/competitions/rsna-breast-cancer-detection/discussion/370045", "author_name": "stpete_ishii", "post_date": "2022-12-02T15:39:54.645000", "votes": 22, "comment_count": 6, "views": 0, "content": "

There are two types of Dicom images in dataset.

\n

TransferSyntaxUID: 1.2.840.10008.1.2.4.90
\nTransferSyntaxDescription: JPEG 2000 Image Compression (Lossless Only)

\n

TransferSyntaxUID: 1.2.840.10008.1.2.4.70
\nTransferSyntaxDescription: JPEG Lossless, Non-Hierarchical, First-Order Prediction (Process 14 [Selection Value 1]): Default Transfer Syntax for Lossless JPEG Image Compression

\n

The image of 1.2.840.10008.1.2.4.90 was able to be processed with the following script, but 1.2.840.10008.1.2.4.70 was not.

\n
 pydicom  dicom\n pydicom  dcmread\nds = dicom.dcmread(path,force=)\narr = ds.pixel_array\n
\n

By installing the followings, both became able to read.

\n

!pip install -U pylibjpeg pylibjpeg-openjpeg pylibjpeg-libjpeg pydicom python-gdcm

", "messages": [ { "id": 2052876, "postDate": "2022-12-02T15:39:54.647Z", "content": "

There are two types of Dicom images in dataset.

\n

TransferSyntaxUID: 1.2.840.10008.1.2.4.90
\nTransferSyntaxDescription: JPEG 2000 Image Compression (Lossless Only)

\n

TransferSyntaxUID: 1.2.840.10008.1.2.4.70
\nTransferSyntaxDescription: JPEG Lossless, Non-Hierarchical, First-Order Prediction (Process 14 [Selection Value 1]): Default Transfer Syntax for Lossless JPEG Image Compression

\n

The image of 1.2.840.10008.1.2.4.90 was able to be processed with the following script, but 1.2.840.10008.1.2.4.70 was not.

\n
 pydicom  dicom\n pydicom  dcmread\nds = dicom.dcmread(path,force=)\narr = ds.pixel_array\n
\n

By installing the followings, both became able to read.

\n

!pip install -U pylibjpeg pylibjpeg-openjpeg pylibjpeg-libjpeg pydicom python-gdcm

", "rawMarkdown": "There are two types of Dicom images in dataset.\n\n**TransferSyntaxUID: 1.2.840.10008.1.2.4.90**\nTransferSyntaxDescription: JPEG 2000 Image Compression (Lossless Only)\n\n**TransferSyntaxUID: 1.2.840.10008.1.2.4.70**\nTransferSyntaxDescription: JPEG Lossless, Non-Hierarchical, First-Order Prediction (Process 14 [Selection Value 1]): Default Transfer Syntax for Lossless JPEG Image Compression\n\nThe image of 1.2.840.10008.1.2.4.90 was able to be processed with the following script, but 1.2.840.10008.1.2.4.70 was not.\n\n```python\nimport pydicom as dicom\nfrom pydicom import dcmread\nds = dicom.dcmread(path,force=True)\narr = ds.pixel_array\n```\n\nBy installing the followings, both became able to read.\n\n`!pip install -U pylibjpeg pylibjpeg-openjpeg pylibjpeg-libjpeg pydicom python-gdcm`", "votes": 20 }, { "id": 2062084, "postDate": "2022-12-11T17:59:17.893Z", "content": "

For train data, we were able to read dicom images using this method under the environment of internet on. However, in this competition, the submission must be done in another notebook under internet off environment and the notebook need to read and analyze unknown test data. So we need to prepare these libraries also under the internet off environment.

", "rawMarkdown": "For train data, we were able to read dicom images using this method under the environment of internet on. However, in this competition, the submission must be done in another notebook under internet off environment and the notebook need to read and analyze unknown test data. So we need to prepare these libraries also under the internet off environment. ", "replies": [ { "id": 2062412, "postDate": "2022-12-12T04:58:22.857Z", "content": "

The way I've done this in previous contests (and hope to do so in this one) is basically to download the necessary packages and their dependencies to my own machine using 'pip download pkg-names', then upload them as datasets to add to my notebook, then run the necessary 'pip install' command inside my notebook code, specifying the filepaths to the uploaded packages.

\n

Hope this helps..

", "rawMarkdown": "The way I've done this in previous contests (and hope to do so in this one) is basically to download the necessary packages and their dependencies to my own machine using 'pip download pkg-names', then upload them as datasets to add to my notebook, then run the necessary 'pip install' command inside my notebook code, specifying the filepaths to the uploaded packages.\n\nHope this helps..", "votes": 1, "replies": [ { "id": 2076572, "postDate": "2022-12-26T16:00:37.667Z", "content": "

Hi. Could you explain how to download the necessary packages and dependencies to my own machine? I installed online and it was ok, but installing these libraries offline hasn't worked.

", "rawMarkdown": "Hi. Could you explain how to download the necessary packages and dependencies to my own machine? I installed online and it was ok, but installing these libraries offline hasn't worked.", "votes": 1 }, { "id": 2079135, "postDate": "2022-12-29T03:26:51.887Z", "content": "

Hi. Isn't this what you need?
\nhttps://www.kaggle.com/datasets/stpeteishii/read-dicom-set

", "rawMarkdown": "Hi. Isn't this what you need?\nhttps://www.kaggle.com/datasets/stpeteishii/read-dicom-set", "votes": 1 }, { "id": 2079387, "postDate": "2022-12-29T09:22:08.423Z", "content": "

Yes, it is! You had already posted it.. Sorry! 😅

", "rawMarkdown": "Yes, it is! You had already posted it.. Sorry! 😅" } ] }, { "id": 2063048, "postDate": "2022-12-12T15:37:59.217Z", "content": "

Thanks for your comment. I prepared a dataset of libraries for reading dicom files. I can install them under the internet-off environment. So, this may be useful for the notebook for submission.
\nhttps://www.kaggle.com/datasets/stpeteishii/read-dicom-set

\n

When use dicomsdl, some difference from pydicom in usage exists.

\n
 dicomsdl\nds = dicomsdl.(path) \narr = ds.pixelData()\n
", "rawMarkdown": "Thanks for your comment. I prepared a dataset of libraries for reading dicom files. I can install them under the internet-off environment. So, this may be useful for the notebook for submission.\nhttps://www.kaggle.com/datasets/stpeteishii/read-dicom-set\n\nWhen use dicomsdl, some difference from pydicom in usage exists.\n```python\nimport dicomsdl\nds = dicomsdl.open(path) \narr = ds.pixelData()\n```\n\n", "votes": 2 } ] } ], "comments": [ { "id": 2062084, "author_name": "stpete_ishii", "author_url": "", "post_date": "2022-12-11T17:59:17.893000", "content": "

For train data, we were able to read dicom images using this method under the environment of internet on. However, in this competition, the submission must be done in another notebook under internet off environment and the notebook need to read and analyze unknown test data. So we need to prepare these libraries also under the internet off environment.

", "votes": 0, "replies": [ { "id": 2062412, "author_name": "David J. Slate", "author_url": "", "post_date": "2022-12-12T04:58:22.857000", "content": "

The way I've done this in previous contests (and hope to do so in this one) is basically to download the necessary packages and their dependencies to my own machine using 'pip download pkg-names', then upload them as datasets to add to my notebook, then run the necessary 'pip install' command inside my notebook code, specifying the filepaths to the uploaded packages.

\n

Hope this helps..

", "votes": 1, "replies": [ { "id": 2076572, "author_name": "Camila Souza", "author_url": "", "post_date": "2022-12-26T16:00:37.667000", "content": "

Hi. Could you explain how to download the necessary packages and dependencies to my own machine? I installed online and it was ok, but installing these libraries offline hasn't worked.

", "votes": 1, "replies": [] }, { "id": 2079135, "author_name": "stpete_ishii", "author_url": "", "post_date": "2022-12-29T03:26:51.887000", "content": "

Hi. Isn't this what you need?
\nhttps://www.kaggle.com/datasets/stpeteishii/read-dicom-set

", "votes": 1, "replies": [] }, { "id": 2079387, "author_name": "Camila Souza", "author_url": "", "post_date": "2022-12-29T09:22:08.423000", "content": "

Yes, it is! You had already posted it.. Sorry! 😅

", "votes": 0, "replies": [] } ] }, { "id": 2063048, "author_name": "stpete_ishii", "author_url": "", "post_date": "2022-12-12T15:37:59.217000", "content": "

Thanks for your comment. I prepared a dataset of libraries for reading dicom files. I can install them under the internet-off environment. So, this may be useful for the notebook for submission.
\nhttps://www.kaggle.com/datasets/stpeteishii/read-dicom-set

\n

When use dicomsdl, some difference from pydicom in usage exists.

\n
 dicomsdl\nds = dicomsdl.(path) \narr = ds.pixelData()\n
", "votes": 2, "replies": [] } ] } ], "raw_markdown_by_id": { "2052876": "There are two types of Dicom images in dataset.\n\n**TransferSyntaxUID: 1.2.840.10008.1.2.4.90**\nTransferSyntaxDescription: JPEG 2000 Image Compression (Lossless Only)\n\n**TransferSyntaxUID: 1.2.840.10008.1.2.4.70**\nTransferSyntaxDescription: JPEG Lossless, Non-Hierarchical, First-Order Prediction (Process 14 [Selection Value 1]): Default Transfer Syntax for Lossless JPEG Image Compression\n\nThe image of 1.2.840.10008.1.2.4.90 was able to be processed with the following script, but 1.2.840.10008.1.2.4.70 was not.\n\n```python\nimport pydicom as dicom\nfrom pydicom import dcmread\nds = dicom.dcmread(path,force=True)\narr = ds.pixel_array\n```\n\nBy installing the followings, both became able to read.\n\n`!pip install -U pylibjpeg pylibjpeg-openjpeg pylibjpeg-libjpeg pydicom python-gdcm`", "2062084": "For train data, we were able to read dicom images using this method under the environment of internet on. However, in this competition, the submission must be done in another notebook under internet off environment and the notebook need to read and analyze unknown test data. So we need to prepare these libraries also under the internet off environment. " } }