basic convolution block Args: inputs: input to the conv layer conv_type: type of convolution. "conv" for Conv2D and "ds" for

bottleneck block for residual convolution Args: inputs: input to the bottleneck residual block filters: number of filters

dict to string for printing options. Args: opt (dict): Option dict. indent_level (int): Indent level. Default: 1. Return:

Helper functionn to create encoder of the model

helper function to create model from given configurations

Parse option file. Args: opt_path (str): Option file path. is_train (str): Indicate whether in training or not. Default: True.

residual convolution bottleneck block Args: inputs: input to the bottleneck residual block filters: number of filters

build encoder, decoder and final model

build encoder and final model

Helper function to create patches Args: images_list: list of image paths masks_list: list of mask paths patch_size: i

Feature Extractor module Args: input_layer: input to the Feature Extractor module Returns: fe_layer4: output of Feature E

Feaure fusion module between learning module and feature extractor module outputs: Args: learning_layer: output of learning module

Get the final model from the given arguments

Helper function for logging

Learning module Args: input_layer: input to the leaning module Returns: learning_layer3: output of learning module

main function for segmentation training

Main function for creating patches and splitting data

main function which creates mask Args: source_image_dir: image directory containing input images source_annotation_dir: annot

Support OrderedDict for yaml. Returns: yaml Loader and Dumper.

Helper function to parse config

prediction function

Helper function which creates masks and croops Args: source_image_dir: image directory containing input images source_annotat

pyramid pooling block from PSPNet Args: input_tensor: input to the pyramid pooling block bin_sizes: bin used to create pyrami

residual convolution block

Helper function to move patches and crop data into train and val

Texture Enhancement Module Args: inputs: input to the tem module Returns: permute_2: output from the tem module

upsampling convolution block

Args: opt: training options

Args: opt: training options

forward pass function