Pyradiomics Customization [CDM]
0 likes274 views
This document discusses how to customize options when extracting radiomic features using the Pyradiomics package in Python. It covers reading image and mask files using SimpleITK, extracting features using a feature extractor object, and customizing options at the feature extractor level and filter class level, including specifying which features to extract, image preprocessing settings, and filter-specific settings.
![Customizing Pyradiomics Options
Yonsei University SeveranceHospital CCIDS
(1) Specifying which features(class) to extract
- Available feature classes are : first order, shape, glcm, glrlm, glszm, gldm, ngtdm
f e a t u r e s _ t o _ e x t r a c t = [ ' f i r s t o r d e r ' , ' s h a p e ' , ' g l c m ' , ' g l r l m ' , ' g l s z m ' , ' g l d m ' , ‘ n g t d m ’ ]
f o r f e a t u r e i n f e a t u r e s _ t o _ e x t r a c t :
e x t r a c t o r. e n a b l e F e a t u r e C l a s s B y N a m e ( f e a t u r e )
https://pyradiomics.readthedocs.io/en/latest/customization.html](https://image.slidesharecdn.com/pyradiomicscustomization20200708cdm-200715042837/85/Pyradiomics-Customization-CDM-10-320.jpg)
![Customizing Pyradiomics Options
Yonsei University SeveranceHospital CCIDS
(2) Specifying settings, which control the pre-processing and customize the behavior of
filters and feature classes
s e t t i n g s = { }
s e t t i n g s [ s e t t i n g _ 1 ] = v a l u e _ 1
s e t t i n g s [ s e t t i n g _ 2 ] = v a l u e _ 2
e x t r a c t o r = r a d i o m i c s . f e a t u r e e x t r a c t o r. R a d i o m i c s F e a t u r e s E x t r a c t o r ( * * s e t t i n g s )
…
Settings are stored in the settings dictionary in the feature extractor class instance,
where the key is the case sensitive setting name.](https://image.slidesharecdn.com/pyradiomicscustomization20200708cdm-200715042837/85/Pyradiomics-Customization-CDM-11-320.jpg)
![Customizing Pyradiomics Options
Yonsei University SeveranceHospital CCIDS
(2) Specifying settings, which control the pre-processing and customize the behavior of
filters and feature classes
s e t t i n g s = { }
s e t t i n g s [ ' n o r m a l i z e ' ] = Tr u e
s e t t i n g s [ ' n o r m a l i z e S c a l e ' ] = 1
s e t t i n g s [ ' r e m o v e O u t l i e r s ' ] = F a l s e
1. Feature Extractor level : Image Normalization](https://image.slidesharecdn.com/pyradiomicscustomization20200708cdm-200715042837/85/Pyradiomics-Customization-CDM-12-320.jpg)
![Customizing Pyradiomics Options
Yonsei University SeveranceHospital CCIDS
(2) Specifying settings, which control the pre-processing and customize the behavior of
filters and feature classes
s e t t i n g s = { }
s e t t i n g s [ ' r e s a m p l e d P i x e l S p a c i n g ' ] = ( 1 , 1 , 1 )
s e t t i n g s [ ' i n t e r p o l a t o r ' ] = ' s i t k B S p l i n e '
1. Feature Extractor level : Resampling the image
Set the size of the voxel in (x, y, z) plane where resampling](https://image.slidesharecdn.com/pyradiomicscustomization20200708cdm-200715042837/85/Pyradiomics-Customization-CDM-13-320.jpg)
![Customizing Pyradiomics Options
Yonsei University SeveranceHospital CCIDS
(2) Specifying settings, which control the pre-processing and customize the behavior of
filters and feature classes
s e t t i n g s = { }
s e t t i n g s [ ' b i n W i d t h ' ] = 2 5
s e t t i n g s [ ' b i n C o u n t ' ] = N o n e
2. Filter class level : Image discretization
Size of bins when making a histogram and for discretization of the image gray level
Specify the number of bins to create. The width of the bin is then determined by the range of in the ROI
* Tips. No definitive evidence on which method is available. Authors advise a fixed bin width.](https://image.slidesharecdn.com/pyradiomicscustomization20200708cdm-200715042837/85/Pyradiomics-Customization-CDM-15-320.jpg)
![Customizing Pyradiomics Options
Yonsei University SeveranceHospital CCIDS
(2) Specifying settings, which control the pre-processing and customize the behavior of
filters and feature classes
s e t t i n g s = { }
s e t t i n g s [ ' d i s t a n c e ' ] = [ [ 1 ] ]
2. Filter class level : Distance to neighbor
Specifies the distance between the center voxel and the neighbor, for which
angles should be generated. (Only valid for GLCM and NGTDM)](https://image.slidesharecdn.com/pyradiomicscustomization20200708cdm-200715042837/85/Pyradiomics-Customization-CDM-16-320.jpg)
![Customizing Pyradiomics Options
Yonsei University SeveranceHospital CCIDS
(2) Specifying settings, which control the pre-processing and customize the behavior of
filters and feature classes
s e t t i n g s = { }
s e t t i n g s [ ' w e i g h t i n g N o r m ' ] = N o n e
2. Filter class level : Texture matric weighting
Which norm should be used when applying distance weighting
(Only valid for GLCM and GLRLM)
• 'manhattan'
• 'euclidian'
• 'infinity'
• 'no_weighting ' : GLCMs are weighted by factor 1 and summed
• None : Applies no weights, mean of values are returned](https://image.slidesharecdn.com/pyradiomicscustomization20200708cdm-200715042837/85/Pyradiomics-Customization-CDM-17-320.jpg)
Pyradiomics Customization [CDM]
- 1. CCIDS 최동민 Yonsei University SeveranceHospital CCIDS PyradiomicsCustomization
- 2. Contents 1. About Pyradiomics 2. Read image and mask files using SimpleITK 3. How to extract radiomic features using Pyradiomics 4. Customizing Pyradiomics options Yonsei University SeveranceHospital CCIDS
- 3. About Pyradiomics Yonsei University SeveranceHospital CCIDS
- 4. About Pyradiomics Yonsei University SeveranceHospital CCIDS - An open-source python package for the extraction of Radiomics features from medical images - Aim to establish a reference standard for Radiomic analysis - Used to calculate single values per feature for a region of interest (“segment-based”) - https://www.radiomics.io/pyradiomics.html
- 5. Read image and mask files using SimpleITK Yonsei University SeveranceHospital CCIDS
- 6. Read image and mask files using SimpleITK Yonsei University SeveranceHospital CCIDS 1. Install SimpleITK package using pip ! p i p i n s t a l l S i m p l e I T K 2. Import SimpleITK i m p o r t S i m p l e I T K a s s i t k 3. Specify image and mask file path i m a g e _ p a t h = ' i m a g e _ s a m p l e _ p a t h . n i i . g z ’ m a s k _ p a t h = ‘ m a s k _ s a m p l e _ p a t h . n i i . g z ' 4. Read image and mask file as SimpleITK object i m a g e = s i t k . R e a d I m a g e ( i m a g e _ p a t h ) m a s k = s i t k . R e a d I m a g e ( m a s k _ p a t h ) p r i n t ( i m a g e ) >>> …
- 7. How to extract radiomic features using Pyradiomics Yonsei University SeveranceHospital CCIDS
- 8. How to extract radiomic features using Pyradiomics Yonsei University SeveranceHospital CCIDS 1. Install Pyradiomics package using pip ! p i p i n s t a l l p y r a d i o m i c s 2. Import Pyradiomics i m p o r t r a d i o m i c s 3. Build feature extractor e x t r a c t o r = r a d i o m i c s . f e a t u r e e x t r a c t o r. R a d i o m i c s F e a t u r e s E x t r a c t o r ( ) 4. Extract features using feature extractor f e a t u r e Ve c t o r = e x t r a c t o r. e x e c u t e ( i m a g e , m a s k )
- 9. Customizing Pyradiomics Options Yonsei University SeveranceHospital CCIDS
- 10. Customizing Pyradiomics Options Yonsei University SeveranceHospital CCIDS (1) Specifying which features(class) to extract - Available feature classes are : first order, shape, glcm, glrlm, glszm, gldm, ngtdm f e a t u r e s _ t o _ e x t r a c t = [ ' f i r s t o r d e r ' , ' s h a p e ' , ' g l c m ' , ' g l r l m ' , ' g l s z m ' , ' g l d m ' , ‘ n g t d m ’ ] f o r f e a t u r e i n f e a t u r e s _ t o _ e x t r a c t : e x t r a c t o r. e n a b l e F e a t u r e C l a s s B y N a m e ( f e a t u r e ) https://pyradiomics.readthedocs.io/en/latest/customization.html
- 11. Customizing Pyradiomics Options Yonsei University SeveranceHospital CCIDS (2) Specifying settings, which control the pre-processing and customize the behavior of filters and feature classes s e t t i n g s = { } s e t t i n g s [ s e t t i n g _ 1 ] = v a l u e _ 1 s e t t i n g s [ s e t t i n g _ 2 ] = v a l u e _ 2 e x t r a c t o r = r a d i o m i c s . f e a t u r e e x t r a c t o r. R a d i o m i c s F e a t u r e s E x t r a c t o r ( * * s e t t i n g s ) … Settings are stored in the settings dictionary in the feature extractor class instance, where the key is the case sensitive setting name.
- 12. Customizing Pyradiomics Options Yonsei University SeveranceHospital CCIDS (2) Specifying settings, which control the pre-processing and customize the behavior of filters and feature classes s e t t i n g s = { } s e t t i n g s [ ' n o r m a l i z e ' ] = Tr u e s e t t i n g s [ ' n o r m a l i z e S c a l e ' ] = 1 s e t t i n g s [ ' r e m o v e O u t l i e r s ' ] = F a l s e 1. Feature Extractor level : Image Normalization
- 13. Customizing Pyradiomics Options Yonsei University SeveranceHospital CCIDS (2) Specifying settings, which control the pre-processing and customize the behavior of filters and feature classes s e t t i n g s = { } s e t t i n g s [ ' r e s a m p l e d P i x e l S p a c i n g ' ] = ( 1 , 1 , 1 ) s e t t i n g s [ ' i n t e r p o l a t o r ' ] = ' s i t k B S p l i n e ' 1. Feature Extractor level : Resampling the image Set the size of the voxel in (x, y, z) plane where resampling
- 14. Customizing Pyradiomics Options Yonsei University SeveranceHospital CCIDS (2) Specifying settings, which control the pre-processing and customize the behavior of filters and feature classes 1. Feature Extractor level : • Pre-cropping • Resegmentation • Mask validation • Miscellaneous
- 15. Customizing Pyradiomics Options Yonsei University SeveranceHospital CCIDS (2) Specifying settings, which control the pre-processing and customize the behavior of filters and feature classes s e t t i n g s = { } s e t t i n g s [ ' b i n W i d t h ' ] = 2 5 s e t t i n g s [ ' b i n C o u n t ' ] = N o n e 2. Filter class level : Image discretization Size of bins when making a histogram and for discretization of the image gray level Specify the number of bins to create. The width of the bin is then determined by the range of in the ROI * Tips. No definitive evidence on which method is available. Authors advise a fixed bin width.
- 16. Customizing Pyradiomics Options Yonsei University SeveranceHospital CCIDS (2) Specifying settings, which control the pre-processing and customize the behavior of filters and feature classes s e t t i n g s = { } s e t t i n g s [ ' d i s t a n c e ' ] = [ [ 1 ] ] 2. Filter class level : Distance to neighbor Specifies the distance between the center voxel and the neighbor, for which angles should be generated. (Only valid for GLCM and NGTDM)
- 17. Customizing Pyradiomics Options Yonsei University SeveranceHospital CCIDS (2) Specifying settings, which control the pre-processing and customize the behavior of filters and feature classes s e t t i n g s = { } s e t t i n g s [ ' w e i g h t i n g N o r m ' ] = N o n e 2. Filter class level : Texture matric weighting Which norm should be used when applying distance weighting (Only valid for GLCM and GLRLM) • 'manhattan' • 'euclidian' • 'infinity' • 'no_weighting ' : GLCMs are weighted by factor 1 and summed • None : Applies no weights, mean of values are returned
- 18. Customizing Pyradiomics Options Yonsei University SeveranceHospital CCIDS (2) Specifying settings, which control the pre-processing and customize the behavior of filters and feature classes 2. Filter class level : • Forced 2D extraction
- 19. Thankyou