如何使用 DALL·E 和 Segment Anything 创建动态蒙版

!pip install torch torchvision torchaudio
!pip install git+https://github.com/facebookresearch/segment-anything.git
!pip install opencv-python pycocotools matplotlib onnxruntime onnx
!pip install requests
!pip install openai
!pip install numpy

!wget https://dl.fbaipublicfiles.com/segment_anything/sam_vit_h_4b8939.pth

import cv2
import matplotlib.pyplot as plt
import matplotlib.image as mpimg
from matplotlib import rcParams
import numpy as np
from openai import OpenAI
import os
from PIL import Image
import requests
from segment_anything import sam_model_registry, SamAutomaticMaskGenerator, SamPredictor
import torch

# Set directories for generation images and edit images
base_image_dir = os.path.join("images", "01_generations")
mask_dir = os.path.join("images", "02_masks")
edit_image_dir = os.path.join("images", "03_edits")

# Point to your downloaded SAM model
sam_model_filepath = "./sam_vit_h_4b8939.pth"

# Initiate SAM model
sam = sam_model_registry["default"](checkpoint=sam_model_filepath)

# Initiate openAI client
client = OpenAI(api_key=os.environ.get("OPENAI_API_KEY", "<your OpenAI API key if not set as env var>"))

def process_dalle_images(response, filename, image_dir):
    # save the images
    urls = [datum.url for datum in response.data]  # extract URLs
    images = [requests.get(url).content for url in urls]  # download images
    image_names = [f"{filename}_{i + 1}.png" for i in range(len(images))]  # create names
    filepaths = [os.path.join(image_dir, name) for name in image_names]  # create filepaths
    for image, filepath in zip(images, filepaths):  # loop through the variations
        with open(filepath, "wb") as image_file:  # open the file
            image_file.write(image)  # write the image to the file

    return filepaths

dalle_prompt = '''
Full length, zoomed out photo of our premium Lederhosen-inspired jumpsuit.
Showcase the intricate hand-stitched details and high-quality leather, while highlighting the perfect blend of Austrian heritage and modern fashion.
This piece appeals to a sophisticated, trendsetting audience who appreciates cultural fusion and innovative design.
'''

# Generate your images
generation_response = client.images.generate(
    model = "dall-e-3",
    prompt=dalle_prompt,
    n=3,
    size="1024x1024",
    response_format="url",
)

filepaths = process_dalle_images(generation_response, "generation", base_image_dir)

# print the new generations
for filepath in filepaths:
    print(filepath)
    display(Image.open(filepath))

# Pick one of your generated images
chosen_image = "images/01_generations/generation_2.png"

# Function to display mask using matplotlib
def show_mask(mask, ax):
    color = np.array([30 / 255, 144 / 255, 255 / 255, 0.6])
    h, w = mask.shape[-2:]
    mask_image = mask.reshape(h, w, 1) * color.reshape(1, 1, -1)
    ax.imshow(mask_image)


# Function to display where we've "clicked"
def show_points(coords, labels, ax, marker_size=375):
    pos_points = coords[labels == 1]
    neg_points = coords[labels == 0]
    ax.scatter(
        pos_points[:, 0],
        pos_points[:, 1],
        color="green",
        marker="*",
        s=marker_size,
        edgecolor="white",
        linewidth=1.25,
    )
    ax.scatter(
        neg_points[:, 0],
        neg_points[:, 1],
        color="red",
        marker="*",
        s=marker_size,
        edgecolor="white",
        linewidth=1.25,
    )

# Choose which mask you'd like to use
chosen_mask = masks[1]

# We'll now reverse the mask so that it is clear and everything else is white
chosen_mask = chosen_mask.astype("uint8")
chosen_mask[chosen_mask != 0] = 255
chosen_mask[chosen_mask == 0] = 1
chosen_mask[chosen_mask == 255] = 0
chosen_mask[chosen_mask == 1] = 255

# We'll save this mask for re-use for our edit
new_mask.save(os.path.join(mask_dir, "new_mask.png"))

# edit an image
edit_response = client.images.edit(
    image=open(chosen_image, "rb"),  # from the generation section
    mask=open(os.path.join(mask_dir, "new_mask.png"), "rb"),  # from right above
    prompt="Brilliant leather Lederhosen with a formal look, detailed, intricate, photorealistic",  # provide a prompt to fill the space
    n=3,
    size="1024x1024",
    response_format="url",
)

edit_filepaths = process_dalle_images(edit_response, "edits", edit_image_dir)