Create an average (mean) face from a batch on input images (PNG, JPEG, GIF).
For each image a dlib shape predictor is run and the largest detected face is picked. The mean corrdinate of each of the facial landmarks is calculated and then a peicewise affine tranfsorm / warp is applied to align all of the faces. Lastly the mean of each pixel location is calculated.
##Dependencies
- scikit-image
- numpy
- matplotlib
- dlib
##Usage
usage: MeanFace.py [-h] [-v] [-p PREDICTOR_PATH] image [image ...]
Average faces in images.
positional arguments:
image Image files to process
optional arguments:
-h, --help show this help message and exit
-v, --verbose Verbose output
-p PREDICTOR_PATH, --predictor-path PREDICTOR_PATH
Path to dlib shape predictor