This projects purpose is to cluster flower images.
Dataset is available in kaggle. Download the dataset, unzip it to where .ipynb file is.
In this dataset, there are 20 classes and 603 flower images.
Clustering achieved by 3 steps.
- Extract features from vgg-16 (remove the last relu-dropout-linear layer) for every single image.
- Apply PCA to decrease the number of features from 4096 to 400.
- Apply kmeans algorithm with k = 20
Install the required libraries
$ pip install -r requirements.txtThen run the script cell by cell.
Clusters (each column is a cluster)
Some clusters doesn't have 10 images, that's why this subplot is not filled.
 |
|---|
| clustering results |
 |
|---|
| Cluster 0 |
 |
|---|
| Cluster 1 |
 |
|---|
| Cluster 2 |
 |
|---|
| Cluster 3 |
 |
|---|
| Cluster 4 |
 |
|---|
| Cluster 5 |
 |
|---|
| Cluster 6 |
 |
|---|
| Cluster 7 |
 |
|---|
| Cluster 8 |
 |
|---|
| Cluster 9 |
 |
|---|
| Cluster 10 |
 |
|---|
| Cluster 11 |
 |
|---|
| Cluster 12 |
 |
|---|
| Cluster 13 |
 |
|---|
| Cluster 14 |
 |
|---|
| Cluster 15 |
 |
|---|
| Cluster 16 |
 |
|---|
| Cluster 17 |
 |
|---|
| Cluster 18 |
 |
|---|
| Cluster 19 |