CT image segmentation - DICOM (Breast CT images)

I have over 500 images from cancer imaging archive. All these are CT images of the chest area (breast). I want to use scikit image to determine which part of the CT represent Fatty tissue and which part represents glands tissue. This might be based of gray scale levels or density I’m not sure.

For instance I want to be able to label image 01 as : 10% fatty tissue 80% fibro glandular tissue and 10% other soft tissue. As at onw, I only adjust the contrast but that really isnt enough for what i want to achieve.
Any biomedical image analyses expert to help ? Any uesful libraries ?
I tried to attach samples of the dicom imges but got an error saying it wasnt an allowed format. Thus, you may find the link to the images on my google drive here. I need help :slight_smile: thanks

I would actually suggest that you use Ilastik for this, which lets you interactively label different regions, then uses machine learning to automatically label the rest of the image as well as all subsequent images in a stack. See http://ilastik.org

2 Likes

Thanks Jni, I am going to look into this.