Graph Based Image Segmentation
Classical method for image segmentation using graph based algorithm [1] implementation using python.
References
[1] Felzenszwalb, P.F., Huttenlocher, D.P. Efficient Graph-Based Image Segmentation. International Journal of Computer Vision 59, 167–181 (2004). [https://doi.org/10.1023/B:VISI.0000022288.19776.77]
[2] Uijlings, J.R.R., van de Sande, K.E.A., Gevers, T. et al. Selective Search for Object Recognition. Int J Comput Vis 104, 154–171 (2013). [https://doi.org/10.1007/s11263-013-0620-5]
[3] Minimum Spanning Tree Using Union Disjoint [https://algorithms.tutorialhorizon.com/disjoint-set-union-find-algorithm-union-by-rank-and-path- compression/]
Required Libraries and Tools
- opencv 3.0+
- numpy
- jupyter notebook
Notes
- The method proposed here for segmentation of images runs in pseudo-polynomial time.
- Current implementation in python is slow, generating result takes atleat 10-15 minutes.
- Minimum spanning tree with path compression algorithm used to implement MST and segmentation tree.
Results
 |
|---|
 |
 |