Manual Histogram Equalization in
MATLAB
This document provides the MATLAB code for manual histogram equalization and includes
detailed explanations for each step.
Step 1: Load Image
Load the grayscale image using 'imread' and display it using 'imshow'. Only these two built-
in functions are allowed.
img = imread('low_contrast_image.png');
imshow(img); title('Original Image');
Step 2: Set Image Size Manually
Specify the image size manually instead of using the 'size' function.
rows = 256; % <-- change this to your image height
cols = 256; % <-- change this to your image width
Step 3: Compute Histogram
Initialize a histogram array with 256 bins and count the frequency of each intensity value in
the image.
histogram = zeros(1, 256);
for i = 1:rows
for j = 1:cols
pixel = img(i, j);
histogram(pixel + 1) = histogram(pixel + 1) + 1;
end
end
Step 4: Compute PDF
Calculate the probability of each intensity value by dividing its count by the total number of
pixels.
total_pixels = rows * cols;
pdf = zeros(1, 256);
for i = 1:256
pdf(i) = histogram(i) / total_pixels;
end
Step 5: Compute CDF
Calculate the cumulative distribution function (CDF) to be used for remapping pixel values.
cdf = zeros(1, 256);
cdf(1) = pdf(1);
for i = 2:256
cdf(i) = cdf(i - 1) + pdf(i);
end
Step 6: Create Mapping Using CDF
Map old intensity values to new ones using the CDF, scaling to [0, 255] and rounding
manually.
mapping = zeros(1, 256);
for i = 1:256
x = cdf(i) * 255;
dec = x - fix(x);
if dec >= 0.5
x = fix(x) + 1;
else
x = fix(x);
end
mapping(i) = x;
end
Step 7: Apply Mapping to Image
Create a new image by replacing each pixel with its mapped value from the mapping array.
equalized_img = img;
for i = 1:rows
for j = 1:cols
val = img(i, j);
equalized_img(i, j) = mapping(val + 1);
end
end
imshow(equalized_img); title('Equalized Image');
Step 8: Compute Mean and Std Dev of Original Image
Calculate the mean and standard deviation manually. Use Newton-Raphson method to
approximate square root.
sum_orig = 0;
for i = 1:rows
for j = 1:cols
sum_orig = sum_orig + img(i,j);
end
end
mean_orig = sum_orig / total_pixels;
sum_sq_orig = 0;
for i = 1:rows
for j = 1:cols
diff = img(i,j) - mean_orig;
sum_sq_orig = sum_sq_orig + diff * diff;
end
end
std_orig = sum_sq_orig / total_pixels;
x = std_orig;
for k = 1:10
x = 0.5 * (x + std_orig / x);
end
std_orig = x;
Step 9: Compute Std Dev of Equalized Image
Repeat the mean and standard deviation calculation for the equalized image.
sum_eq = 0;
for i = 1:rows
for j = 1:cols
sum_eq = sum_eq + equalized_img(i,j);
end
end
mean_eq = sum_eq / total_pixels;
sum_sq_eq = 0;
for i = 1:rows
for j = 1:cols
diff = equalized_img(i,j) - mean_eq;
sum_sq_eq = sum_sq_eq + diff * diff;
end
end
x = sum_sq_eq / total_pixels;
for k = 1:10
x = 0.5 * (x + sum_sq_eq / total_pixels / x);
end
std_eq = x;
Step 10: Report Result
Print and compare standard deviations of original and equalized images to assess contrast
improvement.
fprintf('Original Contrast (Std Dev): %.2f\n', std_orig);
fprintf('Equalized Contrast (Std Dev): %.2f\n', std_eq);
if std_eq > std_orig
disp('Contrast improved after histogram equalization.');
else
disp('No contrast improvement.');
end
