Show layers of the ImageContainer

This example shows how to use squidpy.im.ImageContainer.show().

This function is useful to visualize statically different layers of the squidpy.im.ImageContainer class.

See also

• See Crop images with ImageContainer and Smooth an image for additional examples on methods of the squidpy.im.ImageContainer.

import squidpy as sq

Load the Mibitof dataset.

adata = sq.datasets.mibitof()

Out:

  0%|          | 0.00/19.3M [00:00<?, ?B/s]
  0%|          | 24.0k/19.3M [00:00<01:50, 183kB/s]
  0%|          | 56.0k/19.3M [00:00<01:33, 217kB/s]
  1%|          | 120k/19.3M [00:00<00:59, 338kB/s]
  1%|1         | 256k/19.3M [00:00<00:31, 625kB/s]
  3%|2         | 544k/19.3M [00:00<00:16, 1.19MB/s]
  6%|5         | 1.10M/19.3M [00:00<00:08, 2.29MB/s]
 12%|#1        | 2.23M/19.3M [00:00<00:04, 4.41MB/s]
 23%|##3       | 4.50M/19.3M [00:01<00:01, 8.58MB/s]
 37%|###6      | 7.09M/19.3M [00:01<00:01, 12.1MB/s]
 51%|#####1    | 9.91M/19.3M [00:01<00:00, 15.1MB/s]
 65%|######5   | 12.6M/19.3M [00:01<00:00, 16.8MB/s]
 80%|########  | 15.5M/19.3M [00:01<00:00, 18.4MB/s]
 95%|#########4| 18.3M/19.3M [00:01<00:00, 19.4MB/s]
100%|##########| 19.3M/19.3M [00:01<00:00, 11.4MB/s]

We can briefly visualize the data to understand the type of images we have.

sq.pl.spatial_segment(
    adata,
    library_id=["point16", "point23", "point8"],
    seg_cell_id="cell_id",
    color="Cluster",
    library_key="library_id",
    title=["point16", "point23", "point8"],
)

We have three different tissue samples. We also have segmentation masks for each tissue sample. Let’s extract the image from the anndata.AnnData object and create a squidpy.im.ImageContainer object.

imgs = []
for library_id in adata.uns["spatial"].keys():
    img = sq.im.ImageContainer(adata.uns["spatial"][library_id]["images"]["hires"], library_id=library_id)
    img.add_img(adata.uns["spatial"][library_id]["images"]["segmentation"], library_id=library_id, layer="segmentation")
    img["segmentation"].attrs["segmentation"] = True
    imgs.append(img)
img = sq.im.ImageContainer.concat(imgs)

We can visualize each image of the object with squidpy.im.ImageContainer.show().

img.show("image")

squidpy.im.ImageContainer.show() also allows to overlay the results of segmentation.

img.show("image", segmentation_layer="segmentation", segmentation_alpha=0.5)

Estimated memory usage: 330 MB