Lesson 3: Filtering

Gaussian noise

Noise = randn(size(im)) .* sigma;

Output = img + noise;

Solutions

1. Moving average: replace each pixel with an average of all the values in its neighborhood - resuls smoothed image

Averaging assumptions

1. The true value of pixels are similar to the true value of pixels nearby.
2. The noise added to each pixel is done independently.

Noise function is not reversible since we don’t know the noise function so we can’t actually do the subtraction. Or additive noise destroys the information that we can not recover.

1. Weighted moving average:

Uniform weights: show uncharacteristic plots

Non-uniform weights: more smooth