๋ฐ˜์‘ํ˜•

๐Ÿ”ฌ Medical Image/Processing 5

[BART] BART ์„ค์น˜๋ฐฉ๋ฒ•, Python์„ ์ด์šฉํ•œ coil sensitivity estimation

BART๋Š” Martin Uecker (Graz University of Technology), Jon Tamir (UT Austin), and Michael Lustig (UC Berkeley) ์—ฐ๊ตฌ ๊ทธ๋ฃน์ด ๊ณต๋™์œผ๋กœ ๊ฐœ๋ฐœํ•œ MRI Reconstruction์„ ์œ„ํ•œ Toolbox์ด๋‹ค. ํ”„๋กœ๊ทธ๋ž˜๋ฐ ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์™€ CLI ํ”„๋กœ๊ทธ๋žจ์œผ๋กœ ๊ตฌ์„ฑ๋˜์–ด ์žˆ์œผ๋ฉฐ, Parallel imaging๊ณผ compressed sensing์— ์‚ฌ์šฉํ•  ์ˆ˜ ์žˆ๋Š” ๋‹ค์ฐจ์› ๋ฐฐ์—ด ์—ฐ์‚ฐ, Fourier / wavelet ๋ณ€ํ™˜ ๋“ฑ์˜ ๊ธฐ๋Šฅ์„ ์ œ๊ณตํ•œ๋‹ค. (์ถœ์ฒ˜: ๊ณต์‹ ํ™ˆํŽ˜์ด์ง€) 1. Install ์ฐธ๊ณ : https://mrirecon.github.io/bart/installation.html BART Toolbox BART: Installation Fi..

TorchIO๋ฅผ ์ด์šฉํ•œ 3D ์ด๋ฏธ์ง€ transform (preprocessing/augmentation)

TorchIO๋Š” 3D medical image์˜ loading, preprocessing, augmentation, patch-based sampling์„ ์œ„ํ•œ ์˜คํ”ˆ์†Œ์Šค ํŒŒ์ด์ฌ ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์ด๋‹ค. TorchIO PyPI downloads PyPI version Conda version Google Colab notebooks Documentation status Tests status Lint status Coverage status Code quality Maintainability pre-commit Slack Twitter Twitter commits ... torchio.readthedocs.io ์„ค๋ช…์—๋Š” 3D medical image๋ผ๊ณ  ๋˜์–ด ์žˆ์ง€๋งŒ, ์ผ๋ฐ˜ 3D ์ด๋ฏธ์ง€์—๋„ ์‚ฌ์šฉ ๊ฐ€๋Šฅํ•˜๋‹ค. torch..

dicom ํŒŒ์ผ๋“ค์„ nifti ํŒŒ์ผ๋กœ ๋ฐ”๊ฟ”์ฃผ๋Š” Python library 'dicom2nifti' (+์ผ๋ถ€ dicom ํŒŒ์ผ์ด ์ธ์‹์ด ์•ˆ๋˜๋Š” ๋ฌธ์ œ)

ํด๋” ๋‚ด์˜ dicom ํŒŒ์ผ๋“ค์„ nifti ํŒŒ์ผ๋กœ ๋ฐ”๊ฟ”์ฃผ๋Š” ํŒŒ์ด์ฌ ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ์ด๋‹ค. https://github.com/icometrix/dicom2nifti GitHub - icometrix/dicom2nifti Contribute to icometrix/dicom2nifti development by creating an account on GitHub. github.com ์ด๋ ‡๊ฒŒ ํŽธ๋ฆฌํ•œ ๋ผ์ด๋ธŒ๋Ÿฌ๋ฆฌ๊ฐ€ ์žˆ์—ˆ๋‹ค๋‹ˆ.. ์—ญ์‹œ ํŒŒ์ด์ฌ์€ ์ฐพ์œผ๋ฉด ํŽธ๋ฆฌํ•œ ๊ฒƒ๋“ค์ด ๋„ˆ๋ฌด ๋งŽ๋‹ค ๊ทธ๋™์•ˆ ์ผ์ผ์ด ์ „์ฒ˜๋ฆฌํ–ˆ๋˜๊ฑด ๋จธ๋ฆฌ๊ฐ€ ๋ฉ์ฒญํ•˜๋ฉด ์†์ด ๊ณ ์ƒํ•˜๊ธฐ ๋•Œ๋ฌธ.. Install conda: conda install -c conda-forge dicom2nifti pip: pip install dicom2nifti Usage command line..

Pydicom library ์ด์šฉํ•ด dcm ํŒŒ์ผ ๋ถˆ๋Ÿฌ์˜ค๊ธฐ

Pydicom package๋ฅผ ์ด์šฉํ•ด dcm ํŒŒ์ผ์˜ ์ด๋ฏธ์ง€์™€, ํ—ค๋” ์ •๋ณด๋ฅผ ๋ถˆ๋Ÿฌ์˜ฌ ์ˆ˜ ์žˆ๋‹ค. import pydicom dcm_data = pydicom.read_file(file_path) ์ด๋ฏธ์ง€ ๊ฐ€์ ธ์˜ค๊ธฐ - ์ด๋ฏธ์ง€๋ฅผ numpy array ํ˜•ํƒœ๋กœ ๊ฐ€์ ธ์˜จ๋‹ค. img = dcm_data.pixel_array - ์ถ”๊ฐ€๋กœ, CT image์˜ ๊ฒฝ์šฐ๋Š” RescaleSlope, RescaleIntercept ์ •๋ณด๋ฅผ ๊ฐ€์ ธ์™€ pixel intensity ๊ฐ’์„ ์กฐ์ •ํ•ด์ฃผ์–ด์•ผ ํ•œ๋‹ค. - RescaleSlope๋Š” ๊ฐ pixel intensity๊ฐ’์— ๊ณฑํ•ด์ฃผ๋Š” ๊ฐ’, RescaleIntercept๋Š” ๋”ํ•ด์ฃผ๋Š” ๊ฐ’์ด๋ผ๊ณ  ์ƒ๊ฐํ•˜๋ฉด ๋œ๋‹ค. - pydicom์˜ get(attribute, default_value) method๋Š” ํ•ด๋‹น att..

nibabel affine matrix์˜ ์˜๋ฏธ์™€ affine transform ์ ์šฉํ•˜๋Š” ๋ฐฉ๋ฒ•

https://nipy.org/nibabel/coordinate_systems.html Neuroimaging in Python โ€” NiBabel 3.0.2dev documentation Coordinate systems and affines A nibabel (and nipy) image is the association of three things: The image data array: a 3D or 4D array of image data An affine array that tells you the position of the image array data in a reference space. image metadata (dat nipy.org ์œ„์˜ ๋ฌธ์„œ๋ฅผ ์ •๋ฆฌํ•œ ๋‚ด์šฉ. [์š”์•ฝ] - Image..

1
๋ฐ˜์‘ํ˜•