There are many attempts to use remote sensing techniques in medical purposes such as the delineation of brain tumor spreading. How can we do that job efficiently?
Maybe you can manually create a spectral library of the cells that delineate the tumor or are from the tumor, average the spectra of the library to one spectrum and and then apply the Spectral Angle Mapper (you can also use the normalised dot product). The lowest values indicate the highest similarity. After this you could threshold the result image (segmentation) and fill the gaps using mophological dilation. You can also try to unmix the image using spectra from the tumor and healthy cells. The unmixing result for tumor spectra should show you their abundancies (linear fraction of the spectrum in the mixed spectrum you consider).
Hope it helps
Chris
I agree with Christian Rogaß. This is the only way. Create a library spectrum of tumor cells. Find the absorption and reflectance characteristics of the tumor spectrum. The spectral plot will show you the effective spectral range.
- Badges
- Science topic
- Similar topics
- Geoscience
- Remote Sensing