impoisson

Generates Poisson noise from (image) data. It is a raw but fast implementation based on inversion method and Gaussian approximation

Function which generates Poisson noise / Poisson distribution from the data. It is a raw implementation and uses the inversion method for counts smaller than 30, and gaussian approximation for larger values. The code is vectorized and faster then default imnoise(I,'poisson') and poissrnd

J = impoisson(I)

If I is integer like uint8 the counts equal the input values

If I is single precision then the counts are assumed 1e6 x the input value

If I is double precision then the counts are assumed 1e12 x the input value

Example

X = uint16(repmat((1:1000),1000,1));

Y = impoisson(X);

figure, plot(Y(:)); hold on; plot(X(:),'r')

figure, imshow([X,Y],[])