Adds soft positional embedding with learnable projection.

Input: feature maps from the current stage, the segment map from the previous stage Output: the segmentation map from the current stage

Image to Patch Embedding

Slot Attention module.

Slot Attention module.

A basic Swin Transformer layer for one stage. Args: dim (int): Number of input channels. input_resolution (tuple[int]): Input re

r""" Image to Patch Embedding Args: img_size (int): Image size. Default: 224. patch_size (int): Patch token size. Default: 4.

dataloader for polyp segmentation tasks

r""" Swin Transformer Block. Args: dim (int): Number of input channels. input_resolution (tuple[int]): Input resulotion.

r""" Window based multi-head self attention (W-MSA) module with relative position bias. It supports both of shifted and non-shifted window. A

a structure to represent a weighted edge in graph

Gated Convlution layer with activation (default activation:LeakyReLU) Params: same as conv2d Input: The feature from last layer "I" O

a structure to represent a connected, undirected and weighted graph as a collection of edges.

r""" Patch Merging Layer. Args: input_resolution (tuple[int]): Resolution of input feature. dim (int): Number of input channels.

Slot Attention-based auto-encoder for object discovery.

r""" Swin Transformer A PyTorch impl of : `Swin Transformer: Hierarchical Vision Transformer using Shifted Windows` - https://arxiv

A structure to represent a subset for union-find