What types of filters would you like to see implemented/benchmarked in IBSI installment 2?

Martin Vallières

unread,

May 4, 2018, 2:47:12 PM5/4/18

to Image Biomarker Standardisation Initiative

Preparation for IBSI installment 2 is currently ongoing. This installment will concern the standardization of image filters such as wavelets, etc. If there are particular types of filters that you use in your work and that you would like to see implemented/benchmarked by the IBSI, please let us know!

Alex Zwanenburg

unread,

May 7, 2018, 4:14:50 AM5/7/18

to Image Biomarker Standardisation Initiative

These are the filters that we are currently considering:

Laplacian of Gaussian
Laws kernels
Wavelets (separable)

Haar
Daubechies
Coiflet

Wavelets (non-separable)

Gabor
Riesz

Andrey Fedorov

unread,

May 30, 2018, 5:12:14 PM5/30/18

to Image Biomarker Standardisation Initiative

It would be good to have coverage of the binning approach (i.e., number of bins vs bin width). This issue comes up all the time from the users of pyradiomics (e.g., see https://groups.google.com/d/topic/pyradiomics/0Wx9c9UlUvM/discussion and https://github.com/Radiomics/pyradiomics/issues/378).

Martin Vallières

unread,

May 31, 2018, 3:45:11 PM5/31/18

to Image Biomarker Standardisation Initiative

Hi Andrey,

Thank you very much for your input!

The binning approach is actually covered in IBSI installment 1 (submitted to Medical Image Analysis at the beginning of May): https://arxiv.org/abs/1612.07003. Please see section 2.7. Is that what you meant? Please let me know if there is anything else.