Minimum Spanning Tree Image Segmentation

Source	Euclidean	Manhattan

Pseudocode explaining how one can use the disjoint set data structure to improve Kruskal's algorithm for computing the MST, Kruskal's Algorithm and Disjoint Sets

Dependencies

Python 3 PIL Numpy Scipy

installation: pip3 install pillow numpy scipy

Arguments

Required

ifile: input file path

Optional

sigma: [default] 0.5 - Parameter used by gaussian blurring preproccessing. k: [defualt] 1000 - Value for the threshold function. minSize: [defualt] 80 - Minimum component size enforced by post-processing. ofile: [default] 'segmented.png' - Name of the output file.

Usage

(1) only specifying ifile: python main.py 'testImages/xSqrOv.png'