Condition Deep Convolution Generative Adversarial Network

The multi-condition deep convolution generation countermeasure network is applied to the implementation of Tensorflow in clothing design, which is a stable generation countermeasure network.

Prerequisites

• python 3.5+
• Tensorflow 1.31.1
• Scipy
• pillow
• moviepy
• opencv-python

Usage

First, prepare dataset，Here we used the Fashion- Mnist and Clothing-co -Parsing dataset

• Fashion-Mnist
• Clothing-co-Parsing(CCP)

To train a model with dataset:

python train.py --dataset your_train_dataset --input_height=image_h --output_height=new_image_h --train

To test with an existing model:

python train.py --dataset your_test_dataset --input_height=image_h --output_height=new_image_h

Results

Fashion-mnist

Clothing-co-Parsing:

Reference

Thank you very much to Radford et al(DCGAN)

Thank you very much for the code reference(https://github.com/carpedm20/DCGAN-tensorflow)

Author

Li Minqi, Dong Zhaoxiong, Liu zhe, Li ben, Liu shan, Deng Wei, Gong Mengchan