# Convert image to numpy array for pixel manipulation

img_array = np.array(image)

# Define region around the mouth to clean (based on observation)

# These values may need adjustment depending on precise image characteristics

cleaned_img_array = img_array.copy()

# Approximate region: rows 450 to 550, cols 250 to 400 (manual approximation)

# We'll blur this area slightly to reduce visibility of milk residue

y1, y2 = 450, 550

x1, x2 = 250, 400

# Apply a slight blur to the selected region

region = Image.fromarray(cleaned_img_array[y1:y2, x1:x2])

region = region.filter(ImageFilter.GaussianBlur(radius=2))

# Replace cleaned region in the original image

cleaned_img_array[y1:y2, x1:x2] = np.array(region)

# Convert back to PIL image

cleaned_image = Image.fromarray(cleaned_img_array)

# Apply retro-style filter: increase contrast, add warmth, fade effect

# Step 1: Increase contrast

enhancer = ImageEnhance.Contrast(cleaned_image)

contrast_image = enhancer.enhance(1.3)

# Step 2: Add warmth by increasing red and decreasing blue

r, g, b = contrast_image.split()

r = r.point(lambda i: min(255, i + 15))

b = b.point(lambda i: max(0, i - 10))

warm_image = Image.merge("RGB", (r, g, b))

# Step 3: Add a slight faded effect by lowering saturation

enhancer = ImageEnhance.Color(warm_image)

faded_image = enhancer.enhance(0.8)

# Step 4: Add grain by blending with random noise

noise = np.random.normal(0, 15, (faded_image.size[1], faded_image.size[0], 3)).astype(np.uint8)

noise_img = Image.fromarray(np.clip(np.array(faded_image) + noise, 0, 255).astype(np.uint8))

# Final retro image

final_image = noise_img

# Display the result

final_image.show()