lazyslide.tl.spatial_features

spatial_features(wsi, feature_key, method='smoothing', tile_key='tiles', graph_key=None, layer_key='spatial_features')

Integrate spatial tile graph context with vision model features using spatial feature smoothing.

Parameters:

wsiWSIData

The WSIData object.

feature_keystr

The feature key.

methodstr, default: ‘smoothing’

The method used for spatial feature smoothing. Currently, only ‘smoothing’ is supported.

tile_keystr, default: ‘tiles’

The key of the tiles in the shapes slot.

graph_keystr, optional

The graph key. If None, defaults to ‘{tile_key}_graph’.

layer_keystr, default: ‘spatial_features’

The key for the output layer in the feature table.

Returns:

None.

The transformed feature will be added to the spatial_features layer of the feature AnnData.

Examples

>>> import lazyslide as zs
>>> wsi = zs.datasets.sample()
>>> zs.pp.find_tissues(wsi)
>>> zs.pp.tile_tissues(wsi, 256, mpp=0.5)
>>> zs.tl.feature_extraction(wsi, "resnet50")
>>> zs.pp.tile_graph(wsi)
>>> zs.tl.spatial_features(wsi, "resnet50")
>>> wsi["resnet50_tiles"].layers["spatial_features"]