histocartography.visualization.visualization module

Summary

Classes:

BaseGraphVisualization

Base visualization class

BaseImageVisualization

Base visualization class

HACTVisualization

Hierarchical Cell to Tissue visualization class

InstanceImageVisualization

Instance Image Visualization.

OverlayGraphVisualization

class BaseImageVisualization(instance_style: str = 'outline', color: str = 'black', thickness: int = 1, colormap: Optional[str] = None, alpha: float = 0.5, **kwargs)[source]

Bases: histocartography.pipeline.PipelineStep

Base visualization class

__init__(instance_style: str = 'outline', color: str = 'black', thickness: int = 1, colormap: Optional[str] = None, alpha: float = 0.5, **kwargs) → None[source]

Base visualization class constructor

Parameters

  • instance_style (str) – Defines how to represent the instances (when available). Options are ‘fill’, ‘outline’, ‘fill+outline’. Defaults to ‘outline’.

  • color (str) – Matplotlib named color to (fill) or (outline) the instances. Defaults to ‘black’.

  • thickness (int) – Thickness of the instance outline. Defaults to 1.

  • colormap (str) – Colormap to use to map labels to colors. Defaults to None.

  • alpha (float) – Blending of the background image to the instances. Defaults to 0.5.

abstract draw_instances(canvas: numpy.ndarray, instance_map: numpy.ndarray, instance_attributes: dict)[source]

Abstract method that performs drawing of instances on top of the canvas

Parameters

  • canvas (np.ndarray) – Background on top of which the visualization is drawn.

  • instance_map (np.ndarray) – Segmentation mask of instances, brinary or with individual labels for each entity.

  • instance_attributes (dict) – Dictionary of attributes to be applied to instances.

class InstanceImageVisualization(instance_style: str = 'outline', color: str = 'black', thickness: int = 1, colormap: Optional[str] = None, alpha: float = 0.5, **kwargs)[source]

Bases: histocartography.visualization.visualization.BaseImageVisualization

Instance Image Visualization. Generic instance visualization.

draw_instances(canvas: numpy.ndarray, instance_map: Optional[numpy.ndarray] = None, instance_attributes: Optional[dict] = None) → <module ‘PIL.Image’ from ‘/home/travis/conda/envs/histocartography/lib/python3.7/site-packages/PIL/Image.py’>[source]

Drawing of instances on top of the canvas

Parameters

  • canvas (np.ndarray) – Background on top of which the visualization is drawn.

  • instance_map (np.ndarray) – Segmentation mask of instances, brinary or with individual labels for each entity. Defaults to None.

  • instance_attributes (dict) – Dictionary of attributes to be applied to instances. Defaults to None.

Returns

Canvas with visualization.

Return type

viz_canvas (Image)

class BaseGraphVisualization(instance_visualizer: Optional[histocartography.visualization.visualization.BaseImageVisualization] = None, min_max_color_normalize: bool = True, **kwargs)[source]

Bases: histocartography.pipeline.PipelineStep

Base visualization class

__init__(instance_visualizer: Optional[histocartography.visualization.visualization.BaseImageVisualization] = None, min_max_color_normalize: bool = True, **kwargs) → None[source]

Constructor

Parameters

  • instance_visualizer (BaseImageVisualization) – Instance visualization object. Defaults to None.

  • min_max_color_normalize (bool) – If the node/edge values, eg importance scores, should be min/max normalized. Only relevant if node/edge-level colors are provided. Defaults to True.

abstract draw_nodes(draw: <module 'PIL.ImageDraw' from '/home/travis/conda/envs/histocartography/lib/python3.7/site-packages/PIL/ImageDraw.py'>, graph: dgl.graph.DGLGraph, node_attributes: dict)[source]

Draw nodes on the canvas

abstract draw_edges(draw: <module 'PIL.ImageDraw' from '/home/travis/conda/envs/histocartography/lib/python3.7/site-packages/PIL/ImageDraw.py'>, graph: dgl.graph.DGLGraph, edge_attributes: dict)[source]

Draw edges on the canvas

abstract draw_instances(canvas: numpy.ndarray, instance_map: numpy.ndarray, instance_attributes: dict)[source]

Draw instances on the canvas

abstract graph_preprocessing(graph: dgl.graph.DGLGraph)[source]

preprocesses the graph (e.g., to reorganize spatially)

class OverlayGraphVisualization(node_style: str = 'outline', node_color: str = 'yellow', node_radius: int = 5, edge_style: str = 'line', edge_color: str = 'blue', edge_thickness: int = 2, colormap='viridis', show_colormap=False, **kwargs)[source]

Bases: histocartography.visualization.visualization.BaseGraphVisualization

__init__(node_style: str = 'outline', node_color: str = 'yellow', node_radius: int = 5, edge_style: str = 'line', edge_color: str = 'blue', edge_thickness: int = 2, colormap='viridis', show_colormap=False, **kwargs) → None[source]

Overlay graph visualization class. It overlays a graph drawn with PIL on top of an image canvas using the provided instance_visualizer. Nodes outside of the canvas support willbe ignored.

Parameters

  • node_style (str, optional) – Style to represent the nodes. Options are “filled”, “outline” or “filled+outline”. Defaults to “outline”.

  • node_color (str, optional) – Node color. Defaults to “yellow”.

  • node_radius (int, optional) – Node radius. Defaults to 5.

  • edge_style (str, optional) – Edge style. Defaults to “line”.

  • edge_color (str, optional) – Edge color. Defaults to “blue”.

  • edge_thickness (int, optional) – Edge thickness. Defaults to 2.

  • colormap (str, optional) – Matplotlib colormap. Defaults to “viridis”.

graph_preprocessing(graph: dgl.graph.DGLGraph)[source]

preprocesses the graph (e.g., to reorganize spatially)

draw_nodes(draw: <module 'PIL.ImageDraw' from '/home/travis/conda/envs/histocartography/lib/python3.7/site-packages/PIL/ImageDraw.py'>, graph: dgl.graph.DGLGraph, node_attributes: Optional[dict] = None)[source]

Draws the nodes on top of the canvas.

draw_edges(draw: <module 'PIL.ImageDraw' from '/home/travis/conda/envs/histocartography/lib/python3.7/site-packages/PIL/ImageDraw.py'>, graph: dgl.graph.DGLGraph, edge_attributes: Optional[dict] = None)[source]

Draws the edges on top of the canvas.

draw_instances(canvas: numpy.ndarray, instance_map: Optional[numpy.ndarray] = None, instance_attributes: Optional[dict] = None)[source]

Draws instances on the canvas.

class HACTVisualization(cell_visualizer: Optional[histocartography.visualization.visualization.BaseGraphVisualization] = None, tissue_visualizer: Optional[histocartography.visualization.visualization.BaseGraphVisualization] = None, **kwargs)[source]

Bases: histocartography.pipeline.PipelineStep

Hierarchical Cell to Tissue visualization class

__init__(cell_visualizer: Optional[histocartography.visualization.visualization.BaseGraphVisualization] = None, tissue_visualizer: Optional[histocartography.visualization.visualization.BaseGraphVisualization] = None, **kwargs) → None[source]

Constructor

Parameters

Reference

If you use histocartography in your projects, please cite the following:

@inproceedings{pati2021,
    title = {Hierarchical Graph Representations for Digital Pathology},
    author = {Pushpak Pati, Guillaume Jaume, Antonio Foncubierta, Florinda Feroce, Anna Maria Anniciello, Giosuè Scognamiglio, Nadia Brancati, Maryse Fiche, Estelle Dubruc, Daniel Riccio, Maurizio Di Bonito, Giuseppe De Pietro, Gerardo Botti, Jean-Philippe Thiran, Maria Frucci, Orcun Goksel, Maria Gabrani},
    booktitle = {https://arxiv.org/pdf/2102.11057},
    year = {2021}
}