overlay_mask

overlay_mask(image, mask, colormap, alpha=OVERLAY_ALPHA)

An image with a mask tinted onto it.

Parameters

Name	Type	Description	Default
image	np.ndarray	one image, `height x width x channels`, 0-1 floats or 0-255 integers.	required
mask	np.ndarray	a prediction for it - one-hot or probabilities, `height x width x classes` - or an array of class indices.	required
colormap	list[tuple[int, int, int]]	one colour per class, darkest first, as the specification gives it.	required
alpha	float	how much of the tint shows. The default keeps the tissue legible underneath, which is the point of an overlay.	`OVERLAY_ALPHA`

Returns

Name	Type	Description
	np.ndarray	A `height x width x 3` array of 0-255 integers, which is what the R
	np.ndarray	`overlay_mask()` returns too.

Raises

Name	Type	Description
	PlatypusError	if the image cannot be drawn, or the mask does not cover it.

import numpy as np image = np.full((4, 4, 3), 0.5) mask = np.zeros((4, 4, 2)); mask[1:3, 1:3, 1] = 1 shown = overlay_mask(image, mask, [(0, 0, 0), (255, 0, 0)]) shown.shape, shown.dtype ((4, 4, 3), dtype(‘uint8’)) bool((shown[1, 1] != shown[0, 0]).any()) True