Median Filtering

## Name: Jaewon Son
## Date: November 17 2023
## Honor Statement: I have not given or received any unauthorized assistance on this assignment.
## Link: https://youtu.be/dUWBCpk0HeI


import numpy as np
import cv2


def salt_and_pepper(img, p):

    """Add salt-and-pepper noise to the input image.

    Returns:
        np.ndarray: Returns noisy image with salt-and-pepper noise.
    """

    noisy_img = np.copy(img)
    salt_pepper = np.random.rand(*img.shape[:2])   # Generate random values for each pixel in the image
    noisy_img[salt_pepper < p / 2] = 0   # Set pixels to 0
    noisy_img[salt_pepper > 1 - p / 2] = 255   # Set pixels to 255

    return noisy_img


def median_filter(img, W):

    """Apply median filtering to the input image.

    Returns:
        np.ndarray: Returns filtered image after applying median filtering.
    """

    filtered_img = np.copy(img)
    half_W = W // 2   # Calculate half of the window size

    for i in range(half_W, img.shape[0] - half_W):
        for j in range(half_W, img.shape[1] - half_W):
            neighborhood = img[i - half_W : i + half_W + 1, j - half_W : j + half_W + 1]   # Extract the neighborhhod around each pixel
            filtered_img[i, j] = np.median(neighborhood)   # Replace the pixel value with the median value of the neighborhood

    return filtered_img


img = cv2.imread("C:\\Users\\LG\\Desktop\\DEPAUL UNIV\\2023 Fall\\Python Programming\\DSC 430 - Week 10\\Depaul.jpg")   # Load the image
gray_img = img[:, :, 2]   # Extract the blue channel as a grayscale image
noisy_img = salt_and_pepper(gray_img, p = 0.1)   # Add salt-and-pepper noise with probability p
filtered_img = median_filter(noisy_img, W = 3)   # Apply median filtering with window size W

# Display the images
cv2.imshow("Original Image", img)
cv2.imshow("Noisy Image", noisy_img)
cv2.imshow("Filtered Image", filtered_img)
cv2.waitKey(0)
cv2.destroyAllWindows()