HOW to produce final coloful CTSeg?

[xinnvY]

I have reproduced code using docker,final dirout including c[1-6],wc[1-6],mwc[1-6],y~, I want to asked that is all the output imaging?but how i form the colorful CTSeg as you gaved the example. thanks.

[brudfors]

Hi @xinnvY, That colorful image was created with code not available in CTseg. To reproduce, simply stack the c[1-6] (or wc[1-6]) images and do an argmax along the segmentation class dimension. You can then use some visualization function, like imshow in MATLAB.

[xinnvY]

Thank you for your reply, I wonder if you can share your implementation code of this part, I would appreciate greatly！

---- Replied Message ---- | From | Mikael @.> | | Date | 03/12/2024 16:59 | | To | @.> | | Cc | @.>@.> | | Subject | Re: [WCHN/CTseg] HOW to produce final coloful CTSeg? (Issue #27) |

Hi @xinnvY, That colorful image was created with code not available in CTseg. To reproduce, simply stack the c[1-6] (or wc[1-6]) images and do an argmax along the segmentation class dimension. You can then use some visualization function, like imshow in MATLAB.

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: @.***>

[xinnvY]

Hi@brudfors, i have tried to implement the stack of wc[1-6], stacked shape as image shows [181,217,181,6].I then do an argmax by command 'segmentation = np.argmax(stacked_image, axis=-1)',but i cant open correctly segmenting.nii using itk-SNAP.i wonder if i hava made mistakes in the step of do an argmax ,i really dont understand this detail step as you said along the segmentation class dimesion. looking forward to your answer,thanks!

[brudfors]

I am not sure why you want to open segmenting.nii? The below snippet is more or less what I did, where you have to adjust the colour palette to your liking:

import numpy as np
import matplotlib.pyplot as plt

x = np.random.rand(181, 217, 181, 6)  # this should be the segmentations stacked
y = np.argmax(x, axis=-1)
c = plt.imshow(y[..., 4])

[xinnvY]

Thanks! i see. Actually, i just want to implement the segment task of CT brain image.With best wishes!