Do you have a mathematical definition for under- and over-exposure? This would be a good point to start from.
The pixel values of an overexposed picture are too high and the bright areas might be saturated:
RGB = rand(640, 480, 3); % Some test data
OverExp1 = min(1.0, RGB * 1.2); % Scaled
OverExp2 = min(1.0, RGB + 0.1); % Shifted
For scaling the minimum value of 0.0 is still possible, while for shifting the minimum value is lifted also. A combination of scaling and shifting is possible also.
The under exposure is equivalent.
