This project aims to implement the clustering of images by utilizing Spectral Clustering and Affinity Propagation Clustering together with a number of similarity algorithms, like:
- SIFT: Scale-invariant Feature Transform
- SSIM: Structural Similarity Index
- CW-SSIM: Complex Wavelet Structural Similarity Index
- MSE: Mean Squared Error
The best clustering results are selected according to the calculated performance metrics for clustering:
- Silhouette Coefficient
- Completeness Score
- Homogeneity Score
IPython Notebook (.ipynb file) is included for step-by-step execution of the demo application with extra comments.
The project is using OpenCV 3.1, Scikit-Learn, Scikit-Image and PySSIM for image manipulations, similarity measurements and clustering.
All images have been downloaded from the free-of-charge online service Pixabay: https://pixabay.com
Any kind of copyrights, ownership rights or distribution rights have been considered according to the information, which is available on Pixabay.
Please feel free to ask any questions: oleg.v.puzanov@gmail.com