dear all ,
\ni have some questions of dicom files :-
\n1- what information can we collect from dicom file , except pixel array ( get more statistics ) .
\n2- what is Hounsfield Units and what the effect of applying it in image ?
\n3- how to load images and be sure that this image loaded successfully ?
\n4- what is VOI LUT or Windowing Operation what the effect of applying it in image ?
\nany help ?
1 - There isn't much information in the DICOM meta tags for this competition. You'll find pixel and bit-depth data, but nothing useful clinically outside of window width and center tags (which are used for \"windowing\").
\n2 - H.U. is a concept that only applies to CT (or other calibrated ionizing radiation imaging). H.U. is not used in standard mammography.
\n3 - There are several threads here on extracting pixels most efficiently. There are a few libraries that can do it. Most people use PyDicom or dicomsdl. You will need to use a dataset containing GDCM and PyLIbJPEG to decode some of the files in this comp.
\n4 - VOI LUT stands for Value of Interest Look Up Table. This is a way of mapping 16 bit pixel values to 8 bit values. It's a different concept than regular normalization to 0-255 and allows for better visualization of the difference in tissue densities. You can use the apply_voi_lut() function from PyDicom to use it.
** Note There aren't any VOI LUTs in the competition images though, so PyDicom's behavior is to fall back to applying windowing using the default window width and window center DICOM tags.
", "rawMarkdown": "1 - There isn't much information in the DICOM meta tags for this competition. You'll find pixel and bit-depth data, but nothing useful clinically outside of window width and center tags (which are used for \"windowing\").\n\n2 - H.U. is a concept that only applies to CT (or other calibrated ionizing radiation imaging). H.U. is not used in standard mammography.\n\n3 - There are several threads here on extracting pixels most efficiently. There are a few libraries that can do it. Most people use PyDicom or dicomsdl. You will need to use a dataset containing GDCM and PyLIbJPEG to decode some of the files in this comp.\n\n4 - VOI LUT stands for Value of Interest Look Up Table. This is a way of mapping 16 bit pixel values to 8 bit values. It's a different concept than regular normalization to 0-255 and allows for better visualization of the difference in tissue densities. You can use the `apply_voi_lut()` function from PyDicom to use it. \n\n** Note There aren't any VOI LUTs in the competition images though, so PyDicom's behavior is to fall back to applying windowing using the default window width and window center DICOM tags.", "votes": 5, "replies": [ { "id": 2082413, "postDate": "2023-01-01T14:59:52.847Z", "content": "thanks @davidbroberts
", "rawMarkdown": "thanks @davidbroberts " }, { "id": 2106714, "postDate": "2023-01-19T10:16:05.560Z", "content": "thank you for this,
\nI wanted to highlight a connected issue I highlighted, incase you might want to enlight,
\nthanks again.
\n[https://www.kaggle.com/competitions/rsna-breast-cancer-detection/discussion/377208#2106706]
dear all ,
\ni have some questions of dicom files :-
\n1- what information can we collect from dicom file , except pixel array ( get more statistics ) .
\n2- what is Hounsfield Units and what the effect of applying it in image ?
\n3- how to load images and be sure that this image loaded successfully ?
\n4- what is VOI LUT or Windowing Operation what the effect of applying it in image ?
\nany help ?
1 - There isn't much information in the DICOM meta tags for this competition. You'll find pixel and bit-depth data, but nothing useful clinically outside of window width and center tags (which are used for \"windowing\").
\n2 - H.U. is a concept that only applies to CT (or other calibrated ionizing radiation imaging). H.U. is not used in standard mammography.
\n3 - There are several threads here on extracting pixels most efficiently. There are a few libraries that can do it. Most people use PyDicom or dicomsdl. You will need to use a dataset containing GDCM and PyLIbJPEG to decode some of the files in this comp.
\n4 - VOI LUT stands for Value of Interest Look Up Table. This is a way of mapping 16 bit pixel values to 8 bit values. It's a different concept than regular normalization to 0-255 and allows for better visualization of the difference in tissue densities. You can use the apply_voi_lut() function from PyDicom to use it.
** Note There aren't any VOI LUTs in the competition images though, so PyDicom's behavior is to fall back to applying windowing using the default window width and window center DICOM tags.
", "votes": 5, "replies": [ { "id": 2082413, "author_name": "Ahmed___Gaber____", "author_url": "", "post_date": "2023-01-01T14:59:52.847000", "content": "thanks @davidbroberts
", "votes": 0, "replies": [] }, { "id": 2106714, "author_name": "GUNER", "author_url": "", "post_date": "2023-01-19T10:16:05.560000", "content": "thank you for this,
\nI wanted to highlight a connected issue I highlighted, incase you might want to enlight,
\nthanks again.
\n[https://www.kaggle.com/competitions/rsna-breast-cancer-detection/discussion/377208#2106706]