Spaces:
Running
on
Zero
Running
on
Zero
File size: 2,154 Bytes
bd9da36 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 |
from typing import Optional
import numpy as np
from PIL import Image
def show_masks(
image: np.ndarray,
masks: np.ndarray,
scores: Optional[np.ndarray],
alpha: Optional[float] = 0.5,
display_image: Optional[bool] = False,
only_best: Optional[bool] = True,
autogenerated_mask: Optional[bool] = False,
) -> Image.Image:
if scores is not None:
# sort masks by their scores
sorted_ind = np.argsort(scores)[::-1]
masks = masks[sorted_ind]
if autogenerated_mask:
masks = sorted(masks, key=(lambda x: x["area"]), reverse=True)
else:
# get mask dimensions
h, w = masks.shape[-2:]
if display_image:
output_image = Image.fromarray(image)
else:
# create a new blank image to superimpose masks
if autogenerated_mask:
output_image = Image.new(
mode="RGBA",
size=(
masks[0]["segmentation"].shape[0],
masks[0]["segmentation"].shape[1],
),
color=(0, 0, 0),
)
else:
output_image = Image.new(mode="RGBA", size=(w, h), color=(0, 0, 0))
for i, mask in enumerate(masks):
if not autogenerated_mask:
if mask.ndim > 2: # type: ignore
mask = mask.squeeze() # type: ignore
else:
mask = mask["segmentation"]
# Generate a random color with specified alpha value
color = np.concatenate(
(np.random.randint(0, 256, size=3), [int(alpha * 255)]), axis=0
)
# Create an RGBA image for the mask
mask_image = Image.fromarray((mask * 255).astype(np.uint8)).convert("L")
mask_colored = Image.new("RGBA", mask_image.size, tuple(color))
mask_image = Image.composite(
mask_colored, Image.new("RGBA", mask_image.size), mask_image
)
# Overlay mask on the output image
output_image = Image.alpha_composite(output_image, mask_image)
# Exit if specified to only display the best mask
if only_best:
break
return output_image
|