Skip to content

Visualizing WSIs

torchmil.visualize.patches_to_canvas(patches_list, row_array, column_array, patch_size)

Given a list of WSI patches and their corresponding row and column indices, return a canvas with all the patches.

Parameters:

  • patches_list (list) –

    List of WSI patches. Each patch is a numpy array with shape (patch_size, patch_size, 3).

  • row_array (ndarray) –

    Array with the row indices of the patches.

  • column_array (ndarray) –

    Array with the column indices of the patches.

  • patch_size (int) –

    Size of the patches.

Returns:

  • canvas ( ndarray ) –

    Canvas with all the patches. It has shape (max_row*patch_size, max_column*patch_size, 3).

torchmil.visualize.draw_patches_contour(canvas, row_array, column_array, patch_size, contour_prop=0.05)

Given a canvas with WSI patches already drawn, draw a contour around each patch.

Parameters:

  • canvas (ndarray) –

    Canvas with all the patches. It has shape (max_row*patch_size, max_column*patch_size, 3).

  • row_array (ndarray) –

    Array with the row indices of the patches.

  • column_array (ndarray) –

    Array with the column indices of the patches.

  • patch_size (int) –

    Size of the patches.

  • contour_prop (float, default: 0.05 ) –

    Proportion of the patch size that the contour will cover.

Returns:

  • canvas ( ndarray ) –

    Canvas with the contours drawn. It has shape (max_row*patch_size, max_column*patch_size, 3).

torchmil.visualize.draw_heatmap_wsi(canvas, values, patch_size, row_array, col_array, alpha=0.5, max_color=np.array([0.8392156862745098, 0.15294117647058825, 0.1568627450980392]), min_color=np.array([0.17254901960784313, 0.6274509803921569, 0.17254901960784313]))

Given a canvas with WSI patches already drawn, draw a heatmap on top of the patches. This heatmap is defined by values, which should be normalized between 0 and 1.

Parameters:

  • canvas (ndarray) –

    Canvas with all the patches. It has shape (max_row*patch_size, max_column*patch_size, 3).

  • values (ndarray) –

    Array with the values of the heatmap.

  • patch_size (int) –

    Size of the patches.

  • row_array (ndarray) –

    Array with the row indices of the patches.

  • col_array (ndarray) –

    Array with the column indices of the patches.

  • alpha (float, default: 0.5 ) –

    Alpha value of the heatmap.

  • max_color (ndarray, default: array([0.8392156862745098, 0.15294117647058825, 0.1568627450980392]) ) –

    Color of the highest value of the heatmap.

  • min_color (ndarray, default: array([0.17254901960784313, 0.6274509803921569, 0.17254901960784313]) ) –

    Color of the lowest value of the heatmap.

Returns:

  • canvas ( ndarray ) –

    Canvas with the heatmap drawn. It has shape (max_row*patch_size, max_column*patch_size, 3).