-
Notifications
You must be signed in to change notification settings - Fork 6
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
hints to chess extract #11
Comments
Dear User,
As you pointed out we are planning to add this information in the future.
However in the Methods section of the paper you can have some hints.
We can determine some parameters for the extraction, which have to be tuned
according to the region size analysed:
- *Bilateral filter:* used to denoise the image preserving the edges.
It averages the pixels based on their spatial closeness and their
radiometric similarity, which by default are set to a windows size of 3.
The windows size parameter has this flag--windowsize WINDOWSIZE. The
spatial closeness is obtained by the Gaussian function of the Euclidean
distance between two pixels and its standard deviation. The radiometric
similarity uses the Euclidean distance between two color values, which by
default is set to the mean value of the matrix. The radiometric similarity
has this tag --sigma-spatial. Larger values will average bins with larger
differences.
- *Median filter:* used to smooth the image. It scans the image using
a square shaped array with an area automatically computed according to the
matrix size (by default it is 9, to modify it use --size-medianfilter). It
computes the median of the pixels within the square shaped array to smooth
the signal. Higher values will smooth larger figures, while smaller values
will consider subtle signals (i.e. loops).
- *Otsu’s method:* used to binarize the image. The algorithm searches
for a threshold to separate the pixels in two classes, minimizing the
intra-class variance. By default it is calculated using whole matrix values
to be more refined.
- *Morphological closing:* this last filter is used to remove small
dark spots and connect small bright cracks. It helps to remove remaining
noise and to enclose structures. By default CHESS uses a square-shaped
array of 8 bins (to modify it use the tag --closing-square). Higher values
will enclose larger structures, while lower values will preserve smaller or
more punctuated signals.
If you are interested in this image filters you can check them in here:
https://scikit-image.org/
Sincerely,
S
El mié., 28 oct. 2020 a las 18:38, cgirardot (<notifications@github.com>)
escribió:
… Dear,
in the chess extract doc I read we are planning to release a guide to
this in the future.
Would you by any chance have a draft or few recommendations to give away ?
Between the matrix bin size, the window size and the different extract
parameters; it is a lot to test. Any guidance would be greatly appreciated.
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
<#11>, or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ADZPOZGBV3YIFDLDLWLTRZ3SNBJJZANCNFSM4TCUWJAQ>
.
--
*Silvia Galan Martínez - PhD student *
*Centre Nacional d'Anàlisi Genòmica-Centre de Regulació Genòmica (CNAG-CRG)*
*Structural Genomics Dpt.*
Parc Científic de Barcelona – Torre I
Baldiri Reixac, 4
08028 Barcelona
Tel +34 9340 20580
Email silvia.galan@cnag.crg.eu
|
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Dear,
in the chess extract doc I read
we are planning to release a guide to this in the future.
Would you by any chance have a draft or few recommendations to give away ?
Between the matrix bin size, the window size and the different extract parameters; it is a lot to test. Any guidance would be greatly appreciated.
The text was updated successfully, but these errors were encountered: