Rank Face

A deep learning based model to judge the AQ, Appearance Quotient, of faces. (For Chinese Young Girls Only)
Inspired by Face Rank

Inspiration

Face Rank

My Repository is just a reversion of Face Rank. For more details check this fantastic repo.

SCUT-FBP

This Essay along with its dataset gave me great help in modeling and handling training issues.

Installation

apt-get install python-dev python-pip -y
git clone https://github.com/Entropy-xcy/RankFace
cd ./RankFace
pip install -r requirements.txt
apt-get install python-opencv
# for macOS use 'brew install opencv'
# for Windows try the installation tutorial from opencv official website
wget http://entropy-xcy.bid/faceRank.h5

Demo

python main.py girls.jpg

Here is the output

Training

It is highly recommended to train the model yourself. Some accuracy issues may happen if the platform you have is different from the trainer's.

rm ./faceRank.h5
wget http://entropy-xcy.bid/dataset.zip
unzip dataset.zip
rm dataset.zip
# You may change parameters in the script.
python train.py

Launch API Server

A basic webpage or POST API server build with keras It may still work for mobile platforms

pip install werkzeug
pip install flask
# make sure that you already successfully launched the demo before the next step
# The default port is 5000, you may change it as you wish in the code
python API_server.py

Model Summary:

_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
conv2d_1 (Conv2D)            (None, 128, 128, 32)      896       
_________________________________________________________________
activation_1 (Activation)    (None, 128, 128, 32)      0         
_________________________________________________________________
conv2d_2 (Conv2D)            (None, 126, 126, 32)      9248      
_________________________________________________________________
activation_2 (Activation)    (None, 126, 126, 32)      0         
_________________________________________________________________
max_pooling2d_1 (MaxPooling2 (None, 63, 63, 32)        0         
_________________________________________________________________
dropout_1 (Dropout)          (None, 63, 63, 32)        0         
_________________________________________________________________
flatten_1 (Flatten)          (None, 127008)            0         
_________________________________________________________________
dense_1 (Dense)              (None, 128)               16257152  
_________________________________________________________________
activation_3 (Activation)    (None, 128)               0         
_________________________________________________________________
dropout_2 (Dropout)          (None, 128)               0         
_________________________________________________________________
dense_2 (Dense)              (None, 1)                 129       
=================================================================
Total params: 16,267,425
Trainable params: 16,267,425
Non-trainable params: 0
_________________________________________________________________