I am currently working on Bio-medical image segmentation. I am having doubt that, can we find the type of tumor 'Benign' or 'Malignant' using its growth model i.e. by measuring increase in volume of tumor on consecutive days?
Gunjan,
Although proliferation/increase in size could have some effect on "malignancy", it is probably a better idea to look at the borders of the mass e.g. solid borders might indicate a more benign tumor than one that is extending through tissue layers.
@Mohamed: Thanks for advice.I will take care next time.
@Ryan: So in Image processing term, edge detection and thickness is to be considered. right?
If you have the growth data on real tumors with benign/malignant classification, doing even the most primitive ANOVA followed by maybe a t test would answer that. My guess is there is no single factor relationship, but there could well be a correlation.
Hi Gunjan,
I have already seen some paper working on tumor characterization or some prognosis factors using image processing and machine learning. Only the growth model is not enough to properly characterize tumor type. In addition of the answer of Ryan (borders appearance), local intensity appearance is also employed (heterogeneity vs homogeneity ).
Once you have all these tumor characteristics, you can try a machine learning based classifier to decide in future cases the type of your tumors.