I'm going to use the following dataset to do semantic segmentation with U-Net network.

LGG Segmentation Dataset

This dataset contains brain MR images together with manual FLAIR abnormality segmentation masks. The images were obtained from The Cancer Imaging Archive (TCIA). They correspond to 110 patients included in The Cancer Genome Atlas (TCGA) lower-grade glioma collection with at least fluid-attenuated inversion recovery (FLAIR) sequence and genomic cluster data available. Tumor genomic clusters and patient data is provided in data.csv file.

I have found that brain images are in TIFF format, and they are in RGB (with three channels). I have opened one with Gimp:

enter image description here

I don't know if there is any special information on each colour channel because this question is strange (I don't understand it).

Which channel do I have to use if I want to use greyscale images? Or maybe I can convert them into greyscale.

  • $\begingroup$ Maybe its 3 times the same RoI but with different modality; pre contrast, FLAIR and post contrast. I dont think it is an RGB encodage of greyscale because of the treshold we see on your image. $\endgroup$
    – Chopin
    Commented Aug 3, 2020 at 16:42
  • $\begingroup$ Thanks, I think so. How can I know "who is who". $\endgroup$
    – VansFannel
    Commented Aug 3, 2020 at 16:44

1 Answer 1


From the readme of your kaggle link;

All images are provided in .tif format with 3 channels per image. For 101 cases, 3 sequences are available, i.e. pre-contrast, FLAIR, post-contrast (in this order of channels). For 9 cases, post-contrast sequence is missing and for 6 cases, pre-contrast sequence is missing. Missing sequences are replaced with FLAIR sequence to make all images 3-channel. Masks are binary, 1-channel images. They segment FLAIR abnormality present in the FLAIR sequence (available for all cases).

So you could use all three channel.

  • $\begingroup$ Thanks a lot. I need to use FLAIR images, so I think I have to use second channel. $\endgroup$
    – VansFannel
    Commented Aug 4, 2020 at 5:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.