local5's picture
Upload folder using huggingface_hub
452e643 verified
Raw History Blame Contribute Delete
5.6 kB
import gradio as gr
import cv2
import numpy as np
from PIL import Image, ImageDraw, ImageFont, ImageFilter, ImageEnhance
from skimage.metrics import structural_similarity as ssim
import tempfile
# ASCII characters to map pixel intensity
ASCII_CHARS = [
' ', '.', ',', ':', ';', 'i', 'l', '!', 'I', '1', 't', 'o', 'r', 'x', 'z', 'v',
'u', 'n', 'm', 'w', 'Q', 'B', 'N', 'M', '@'
]
# Step 1: Resize image and perform general Sobel edge detection
def preprocess_image(image_path, new_width=150):
# Open the image and convert to grayscale
image = Image.open(image_path)
print(type(image))
# Step 1: Resize the image for ASCII aspect ratio
width, height = image.size
aspect_ratio = height / width * 0.55 # Adjusting for ASCII aspect ratio
new_height = int(aspect_ratio * new_width)
# # Step 3: Show the equalized image
# equalized_image.show(title="Equalized Image")
opencv_image = np.array(image)
clahe = cv2.createCLAHE(clipLimit=2.0, tileGridSize=(8, 8))
equalized_image = clahe.apply(opencv_image)
# Convert to OpenCV format for Sobel edge detection
opencv_image = np.array(equalized_image)
# Step 4: Sobel edge detection
sobelx = cv2.Sobel(opencv_image, cv2.CV_64F, 1, 0, ksize=5) # Horizontal gradients
sobely = cv2.Sobel(opencv_image, cv2.CV_64F, 0, 1, ksize=5) # Vertical gradients
edges = np.hypot(sobelx, sobely) # Combine gradients
# Normalize the result to fit into the 0-255 grayscale range
edges = np.uint8(255 * edges / np.max(edges))
# Step 5: Sharpen the edges
edges_image = Image.fromarray(edges)
edges_image.show(title="Sharpened Edges")
enhancer = ImageEnhance.Contrast(edges_image)
enhanced_edges = enhancer.enhance(3.0) # Increase contrast by a factor (2.0 is an example)
sharpened_edges = enhanced_edges.filter(ImageFilter.SHARPEN)
# sharpened_edges = edges_image.filter(ImageFilter.SHARPEN)
# Step 6: Show the sharpened edges image
sharpened_edges.show(title="Sharpened Edges")
sharpened_edges = sharpened_edges.resize((new_width, new_height))
sharpened_edges.show(title="Sharpened Edges")
return sharpened_edges, sharpened_edges
# Step 2: Convert image to ASCII art
def image_to_ascii(image):
pixels = np.array(image)
ascii_str = ""
for row in pixels:
for pixel in row:
ascii_str += ASCII_CHARS[pixel // 10] # Mapping pixel intensity to ASCII (256 levels / 25 characters)
ascii_str += "\n"
return ascii_str
# Step 3: Convert ASCII text back to image
def ascii_to_image(ascii_art, font_size=12, output_file='ascii_art.png'):
font_path = "DejaVuSansMono.ttf" # Update with correct path if needed
font = ImageFont.truetype(font_path, font_size)
lines = ascii_art.splitlines()
max_width = max(font.getbbox(line)[2] for line in lines) # Get max width
img_height = len(lines) * font_size
image = Image.new("RGB", (max_width, img_height), "white")
draw = ImageDraw.Draw(image)
for y, line in enumerate(lines):
draw.text((0, y * font_size), line, fill="black", font=font)
image.save(output_file)
return image
# Step 4: Calculate Structural Similarity Index (SSIM)
def compare_ssim(image_path_1, image_path_2):
image1 = cv2.imread(image_path_1, cv2.IMREAD_GRAYSCALE)
image2 = cv2.imread(image_path_2, cv2.IMREAD_GRAYSCALE)
# Resize images to the same size if needed
image2 = cv2.resize(image2, (image1.shape[1], image1.shape[0]))
# Compute SSIM between the two images
score, _ = ssim(image1, image2, full=True)
return score
# Usage Example
if __name__ == "__main__":
# Path to input image
input_image_path = "image8a.jpg"
# Step 1: Preprocess and edge detection with Sobel filter
resized_image, edges = preprocess_image(input_image_path)
# Step 2: Convert the edge-detected image to ASCII
ascii_art = image_to_ascii(edges)
print("ASCII Art:\n", ascii_art)
# Step 3: Convert ASCII back to image
ascii_image_path = "ascii_image.jpg"
ascii_img = ascii_to_image(ascii_art, 12, ascii_image_path)
# Step 4: Compare SSIM
ssim_score = compare_ssim(input_image_path, ascii_image_path)
print(f"SSIM between original and ASCII image: {ssim_score}")
# def process_image(image):
# # Step 1: Preprocess and convert to ASCII
# edges, _ = preprocess_image(image)
# ascii_art = image_to_ascii(edges)
# # Step 2: Convert ASCII art back to an image for SSIM calculation
# ascii_image = ascii_to_image(ascii_art)
# # Step 3: Calculate SSIM
# ssim_score = compare_ssim(image, ascii_image)
# # Step 4: Return ASCII art and SSIM score
# with tempfile.NamedTemporaryFile(delete=False, suffix=".txt") as f:
# f.write(ascii_art.encode())
# ascii_txt_path = f.name
# return ascii_art, ssim_score, ascii_txt_path
# # Gradio Interface
# def gradio_interface(image):
# print(type(image))
# ascii_art, ssim_score, ascii_txt_path = process_image(image)
# return ascii_art, ssim_score, ascii_txt_path
# # Setup Gradio interface
# interface = gr.Interface(
# fn=gradio_interface,
# inputs="image",
# outputs=["text", "number", "file"],
# title="Image to ASCII Art with SSIM",
# description="Upload an image, see the ASCII art, the SSIM score, and download the ASCII art as a text file."
# )
interface.launch()