FaceAttribute

Six face attributes (Attractive, EyeGlasses, Male, MouthOpen, Smiling, Young) predication from a single face image.

PyTorch implementation. Trained using celebA dataset.

Dependencies

• Python 3.6+ (Anaconda)
• PyTorch-0.2.0 +
• scipy, numpy, sklearn etc.
• OpenCV3 (Python)

Tested on Ubuntu 14.04 LTS, Python 3.6 (Anaconda), PyTorch-0.3.0, CUDA 8.0, cuDNN 5.0

Usage

Data Preprocessing

detMTCNN_celebA.py

AttrListGen.py

Data loader

dataloadercelebA.py

Model Training

TrainAttrPreRes18V0.py

TrainAttrPreV0.py

Models

AttrPreModelRes18_256V0.py

AttrPreModelRes34_256V0.py

Model evaluation

AttrEvaRes18_256V0.py

AttrEvaRes34_256V0.py

Results

Focal Loss:

Attractive 0.8231

EyeGlasses 0.9980

Male 0.9721

MouthOpen 0.9407

Smiling 0.9200

Young 0.8776