Face Clustering:

This is a simple Python script which uses the images from /dataset directory and generates the output feature vector is 128-d (i.e., a list of 128 real-valued numbers) that is used to quantify the face then group the similar faces together.

Here the /dataset directory contains images (20 each) of 5 Indian Cricketers: Dravid, Ganguly, MS Dhoni, Sachin Tendulkar and Virat Kohli.

1. face_cluster.py : Detects similar faces from given directoy and show them together
2. face_cluster_encodes_colab.pkl : pickle file containing the encodings of images in /dataset

Sample output:

(/dataset Images source: Google Images)

Thanks to Adrian Rosebrock from https://www.pyimagesearch.com/

Link : https://www.pyimagesearch.com/2018/07/09/face-clustering-with-python/