Image Enhancement under Data-Dependent Multiplicative Gamma Noise

Algorithm 1

Image enhancement filter.

(1) Input the noisy image 𝑢₀ corrupted by out-of-focus blur and data-dependent

  multiplicative Gamma noise.

(2) Calculate the regularization weight 𝜆₂ as 0.5/𝜎_𝑛² where 𝜎_𝑛 is the noise

   standard deviation and 𝜆₁=0.02 (this was set empirically).

(3) Specify the stopping threshold value 𝑡.

(4) Select a time step of Δ𝑡=0.1 for the PDE evolution.

(5) Initialize the current restored image 𝑢^𝑛_𝑟=0 and 𝑢^𝑟_𝑛+1=𝑢₀

(6) while  (‖𝑢_𝑟^𝑛+1−𝑢^𝑛_𝑟‖/‖𝑢^𝑛_𝑟‖) >𝑡  do,

(7)   Assign 𝑢^𝑛_𝑟=𝑢_𝑟^𝑛+1

(8)   Evaluate 𝑢_𝑟^𝑛+1 using the PDE defined in (10).

(9) end while