3

I am trying to implement vessel extraction from Retinal fundus image from the paper Multilayered thresholding-based blood vessel segmentationfor screening of diabetic retinopathy: M. Usman Akram • Shoab A. Khan

It requires me to determine 2D Gabor Wavelet of the input image to enhance the blood vessels only. The response of the 2D Gabor wavelet when applied on green channel of the image looks something like this:

enter image description here

I read upon Pywavelets but it provides Wavelet transform for only 1D signals. Even scipy does not have an implementation. There are some Matlab implementations but I don't seem to understand it. I came across this: https://pythonhosted.org/bob.ip.gabor/ but I couldn't comprehend it nor was it installable on Windows

Is there any library that provides me with the said functionality? Or, how can I implement it by hand using Numpy/Scipy?

Thanks

Sagar B Hathwar
  • 536
  • 6
  • 18
  • 1
    Are [openCV Gabor filters](https://cvtuts.wordpress.com/2014/04/27/gabor-filters-a-practical-overview/) not good for what you're trying to do? – Hirabayashi Taro Feb 01 '18 at 04:45
  • @HirabayashiTaro The paper specifically says to use 2D Gabor Wavelet transform for the purpose, comparing it with performance of Gabor Filter. Hence, I specifically need Gabor wavelet transform. I did look at cv.getGaborKernel, but that is not what I need. – Sagar B Hathwar Feb 01 '18 at 04:57

0 Answers0