my research work is done till segmentation using fcm and now i want to check whether the input Brain mri is normal or abnormal and if it is abnormal then whether it is begnin or melignin. suggestion for how to proceed further are awaited.
Different kinds of lesions might be contained in a brain MRI. A brain MRI is abnormal if it contains at least one class of lesion. To my knowledge, It would be a complex problem to tell whether a brain MRI image is normal or not.
If you are only interested in detecting tumor from brain MRI images, then various machine learning techniques can be applied to detect tumor from images. The same techniques can also be applied to tell whether a detected lesion is begnin or malignant.
@lisheng yes sir it is difficult. i am trying by dividing brain mri into two halves from center and then compare its histogram.but m not getting the output properly..
any suggestion??
You can refer to the methods in papers such as these http://www98.griffith.edu.au/dspace/bitstream/handle/10072/52237/81233_1.pdf?sequence=1
@Kartheek thanks for the paper. This gave me an idea where i was getting wrong. i tried dividing mri into two halves and computed their histogram of each hemisphere. then i tried to get difference in their histogram by pdist2 but i am not getting satisfied results. any suggestion where i am getting wrong.
It has certain rationality to judge the abnormality of a brain mri based on the symmetry of the brain.
I think that small patch based comparison might be a suitable comparison mode. Histogram based comparison might not be able to correctly tell differences between corresponding patches.
even, patch baded comparison cannot detect all different kinds of lesions.
How many brain mri images (normal or abnormal ) do you have now?
@LISHENG i have used BRATS 2012 datasets and in that i have around 80 images out of which around 10 are normal brain images.
- Badges
- Science topic
- Similar topics
- Biological Science
- Immunology
- Antibodies