Image segmentation various methods almost from scratch:
in this project i will demonstrate several methods for image segmentation. my goal is to implement each method by myself, and then possibly, add some nice adjustments for improving.
the evaluation method will be F - score measurement which uses precision and recall (specific formulation inside the notebook)
I will use images from The Weizmann Horse Database which contains 328 side-view color images of horses that were also manually segmented for ground truth evaluations. the credit goes to Eran Borenstein and his site: https://www.msri.org/people/members/eranb/
in this project you can find the following implementations:
for automatic seeds generation i found local minimas using peak_local_max from skimage.feature for region growin i used BFS & Queue data structure
Accepting both gray and RGB images
As presented in Normalized Cuts and Image Segmentation,by Shi and Malik. link
i used 0 and median for thresholding the 2nd lowest eigenvalue, and limited the recursion to depth = 3. the partition works as perfect binary tree