Evaluates the precision & recall of AWS Rekognition on the LFW dataset.
To run:
-
Setup a Python 3.6 virtual environment and install the requirements: $ pip - r requirements.txt
-
Download and unzip the LFW images into a folder called 'input_images': http://vis-www.cs.umass.edu/lfw/
-
Put your AWS credentials into process.py
-
$ python process.py
TODOS:
- k-fold cross-validation
- plotting results with plot.ly