sam2-playground / modules /mask_utils.py
jhj0517
Update `invert_masks()`
44f2d47
raw
history blame
8.42 kB
import cv2
import numpy as np
from typing import Dict, List, Tuple
import colorsys
from pytoshop import layers
from pytoshop.enums import BlendMode
from pytoshop.core import PsdFile
from modules.constants import DEFAULT_COLOR, DEFAULT_PIXEL_SIZE
def decode_to_mask(seg: np.ndarray[np.bool_] | np.ndarray[np.uint8]) -> np.ndarray[np.uint8]:
"""Decode to uint8 mask from bool to deal with as images"""
if isinstance(seg, np.ndarray) and seg.dtype == np.bool_:
return seg.astype(np.uint8) * 255
else:
return seg.astype(np.uint8)
def invert_masks(masks: List[Dict]) -> List[Dict]:
"""Invert the masks. Used for background masking"""
inverted = 1 - masks
return inverted
def generate_random_color() -> Tuple[int, int, int]:
"""Generate random color in RGB format"""
h = np.random.randint(0, 360)
s = np.random.randint(70, 100) / 100
v = np.random.randint(70, 100) / 100
r, g, b = colorsys.hsv_to_rgb(h/360, s, v)
return int(r * 255), int(g * 255), int(b * 255)
def create_base_layer(image: np.ndarray) -> List[np.ndarray]:
"""Create a base layer from the image. Used to keep original image"""
rgba_image = cv2.cvtColor(image, cv2.COLOR_RGB2RGBA)
return [rgba_image]
def create_mask_layers(
image: np.ndarray,
masks: List[Dict]
) -> List[np.ndarray]:
"""
Create list of images with mask data. Masks are sorted by area in descending order.
Args:
image: Original image
masks: List of mask data
Returns:
List of RGBA images
"""
layer_list = []
sorted_masks = sorted(masks, key=lambda x: x['area'], reverse=True)
for info in sorted_masks:
rle = info['segmentation']
mask = decode_to_mask(rle)
rgba_image = cv2.cvtColor(image, cv2.COLOR_RGB2RGBA)
rgba_image[..., 3] = cv2.bitwise_and(rgba_image[..., 3], rgba_image[..., 3], mask=mask)
layer_list.append(rgba_image)
return layer_list
def create_mask_gallery(
image: np.ndarray,
masks: List[Dict]
) -> List:
"""
Create list of images with mask data. Masks are sorted by area in descending order. Specially used for gradio
Gallery component. each element has image and label, where label is the part number.
Args:
image: Original image
masks: List of mask data
Returns:
List of [image, label] pairs
"""
mask_array_list = []
label_list = []
sorted_masks = sorted(masks, key=lambda x: x['area'], reverse=True)
for index, info in enumerate(sorted_masks):
rle = info['segmentation']
mask = decode_to_mask(rle)
rgba_image = cv2.cvtColor(image, cv2.COLOR_RGB2RGBA)
rgba_image[..., 3] = cv2.bitwise_and(rgba_image[..., 3], rgba_image[..., 3], mask=mask)
mask_array_list.append(rgba_image)
label_list.append(f'Part {index}')
return [[img, label] for img, label in zip(mask_array_list, label_list)]
def create_mask_combined_images(
image: np.ndarray,
masks: List[Dict]
) -> List:
"""
Create an image with colored masks. Each mask is colored with a random color and blended with the original image.
Args:
image: Original image
masks: List of mask data
Returns:
[image, label] pairs
"""
final_result = np.zeros_like(image)
used_colors = set()
for info in masks:
rle = info['segmentation']
mask = decode_to_mask(rle)
while True:
color = generate_random_color()
if color not in used_colors:
used_colors.add(color)
break
colored_mask = np.zeros_like(image)
colored_mask[mask > 0] = color
blended = cv2.addWeighted(image, 0.3, colored_mask, 0.7, 0)
final_result = np.where(mask[:, :, np.newaxis] > 0, blended, final_result)
combined_image = np.where(final_result != 0, final_result, image)
hsv = cv2.cvtColor(combined_image, cv2.COLOR_BGR2HSV)
hsv[:, :, 1] = np.clip(hsv[:, :, 1] * 1.5, 0, 255)
enhanced = cv2.cvtColor(hsv, cv2.COLOR_HSV2BGR)
return [enhanced, "Masked"]
def create_mask_pixelized_image(
image: np.ndarray,
masks: List[Dict],
pixel_size: int = DEFAULT_PIXEL_SIZE
) -> np.ndarray:
"""
Create a pixelized image with mask.
Args:
image: Original image
masks: List of mask data
pixel_size: Pixel size for pixelization
Returns:
Pixelized image
"""
final_result = image.copy()
def pixelize(img: np.ndarray, mask: np.ndarray[np.uint8], pixel_size: int):
h, w = img.shape[:2]
temp = cv2.resize(img, (w // pixel_size, h // pixel_size), interpolation=cv2.INTER_LINEAR)
pixelated = cv2.resize(temp, (w, h), interpolation=cv2.INTER_NEAREST)
return np.where(mask[:, :, np.newaxis] > 0, pixelated, img)
for info in masks:
rle = info['segmentation']
mask = decode_to_mask(rle)
pixelated_segment = pixelize(final_result, mask, pixel_size)
final_result = np.where(mask[:, :, np.newaxis] > 0, pixelated_segment, final_result)
return final_result
def create_solid_color_mask_image(
image: np.ndarray,
masks: List[Dict],
color_hex: str = DEFAULT_COLOR
) -> np.ndarray:
"""
Create an image with solid color masks.
Args:
image: Original image
masks: List of mask data
color_hex: Hex color code
Returns:
Image with solid color masks
"""
final_result = image.copy()
def hex_to_bgr(hex_color: str):
hex_color = hex_color.lstrip('#')
rgb = tuple(int(hex_color[i:i + 2], 16) for i in (0, 2, 4))
return rgb[::-1]
color_bgr = hex_to_bgr(color_hex)
for info in masks:
rle = info['segmentation']
mask = decode_to_mask(rle)
solid_color_mask = np.full(image.shape, color_bgr, dtype=np.uint8)
final_result = np.where(mask[:, :, np.newaxis] > 0, solid_color_mask, final_result)
return final_result
def insert_psd_layer(
psd: PsdFile,
image_data: np.ndarray,
layer_name: str,
blending_mode: BlendMode
) -> PsdFile:
"""
Insert a layer into the PSD file using pytoshop
Args:
psd: PSD file object from the pytoshop
image_data: Image data
layer_name: Layer name
blending_mode: Blending mode from pytoshop
Returns:
Updated PSD file object
"""
channel_data = [layers.ChannelImageData(image=image_data[:, :, i], compression=1) for i in range(4)]
layer_record = layers.LayerRecord(
channels={-1: channel_data[3], 0: channel_data[0], 1: channel_data[1], 2: channel_data[2]},
top=0, bottom=image_data.shape[0], left=0, right=image_data.shape[1],
blend_mode=blending_mode,
name=layer_name,
opacity=255,
)
psd.layer_and_mask_info.layer_info.layer_records.append(layer_record)
return psd
def save_psd(
input_image_data: np.ndarray,
layer_data: List,
layer_names: List,
blending_modes: List,
output_path: str
):
"""
Save the image with multiple layers as a PSD file
Args:
input_image_data: Original image data
layer_data: List of images to be saved as layers
layer_names: List of layer names
blending_modes: List of blending modes
output_path: Output path for the PSD file
"""
psd_file = PsdFile(num_channels=3, height=input_image_data.shape[0], width=input_image_data.shape[1])
psd_file.layer_and_mask_info.layer_info.layer_records.clear()
for index, layer in enumerate(layer_data):
psd_file = insert_psd_layer(psd_file, layer, layer_names[index], blending_modes[index])
with open(output_path, 'wb') as output_file:
psd_file.write(output_file)
def save_psd_with_masks(
image: np.ndarray,
masks: List[Dict],
output_path: str
):
"""
Save the psd file with masks data.
Args:
image: Original image
masks: List of mask data
output_path: Output path for the PSD file
"""
original_layer = create_base_layer(image)
mask_layers = create_mask_layers(image, masks)
names = [f'Part {i}' for i in range(len(mask_layers))]
modes = [BlendMode.normal] * (len(mask_layers)+1)
save_psd(image, original_layer+mask_layers, ['Original_Image']+names, modes, output_path)