ResNet Auto-Encoder

This repository contains Tensorflow code for an Auto-Encoder architecture built with Residual Blocks.

Default Architecture Parameters:

model = ResNetAE(input_shape=(256, 256, 3),
                 n_ResidualBlock=8,
                 n_levels=4,
                 z_dim=128,
                 bottleneck_dim=128,
                 bUseMultiResSkips=True)

• input_shape: A tuple defining the input image shape for the model
• n_ResidualBlock: Number of Convolutional residual blocks at each resolution
• n_levels: Number of scaling resolutions, at each increased resolution, the image dimension halves and the number of filters channel doubles
• z_dim: Number of latent dim filters
• bottleneck_dim: AE/VAE vectorized latent space dimension
• bUseMultiResSkips: At each resolution, the feature maps are added to the latent/image output (green path in diagram)

Encoder

The encoder expects a 4-D Image Tensor in the form of [Batch x Height x Width x Channels]. The output z would be of shape [Batch x Height/(2**n_levels) x Width/(2**n_levels) x z_dim].

encoder = ResNetEncoder(n_ResidualBlock=8, 
                        n_levels=4,
                        z_dim=10, 
                        bUseMultiResSkips=True)

N.B. It is possible to flatten z by tf.layers.dense for a vectorised latent space, as long as the shape is preserved for the decoder during the unflatten process.

Decoder

The decoder expects a 4-D Feature Tensor in the form of [Batch x Height x Width x Channels]. The output x_out would be of the shape [Batch x Height*(2**n_levels) x Width*(2**n_levels) x output_channels]

decoder = ResNetDecoder(n_ResidualBlock=8, 
                        n_levels=4,
                        output_channels=3, 
                        bUseMultiResSkips=True)

Residual Block

The Residual Block uses the Full pre-activation ResNet Residual block by He et al.

TODO: implementation changed to Conv-Batch-Relu, update figure

If you find this work useful for your research, please cite:

@article{ResNetAE, 
  Title={{R}es{N}et{AE}-https://github.com/farrell236/ResNetAE}, 
  url={https://github.com/farrell236/ResNetAE},  
  Author={Hou, Benjamin}, 
  Year={2019}
}