Models

The models module within our library provides implementations of deep learning models tailored for computational imaging \(\reconnet\). These models leverage the latest advancements in neural networks to offer robust solutions for image reconstruction, enhancement, and segmentation.

List of models

The models module contains the following models:

colibri.models.autoencoder.Autoencoder

Autoencoder Model

colibri.models.unet.Unet

Unet Model

List of custom layers

The custom_layers module within our library provides implementations of custom layers that are used in the models. These layers are designed to enhance the performance of the models by providing additional flexibility and control over the network architecture.

colibri.models.custom_layers.Activation

Activation Layer

colibri.models.custom_layers.convBlock

Convolutional Block

colibri.models.custom_layers.downBlock

Spatial downsampling and then convBlock

colibri.models.custom_layers.upBlock

Spatial upsampling and then convBlock

colibri.models.custom_layers.upBlockNoSkip

Spatial upsampling and then convBlock

colibri.models.custom_layers.outBlock

Convolutional Block with 1x1 kernel and without activation