Wgan For West-African Masks

Implementation of a Wasserstein GAN, for generating images of traditional west African masks, using TensorFlow. The architecture is based on that described in the paper, 'Using WGAN for Improving Imbalanced Classification Performance', by Bhatias and Dahyotr.

Link to paper :

http://ceur-ws.org/Vol-2563/aics_34.pdf

Sample Images Generated:

Run locally :

1.Download and navigate to base directory.

2.Run python3 gan_implementation_for_west_african_masks.py

Run on Google CoLab:

1.Open notebook and ensure tf.version 2.x

2.Upload data_set.zip file and run.

Generated Images At Each Epoch:

Epoch 0

Epoch 50

Epoch 500

Epoch 1000

Epoch 1500