Maximum Likelihood Estimation for an image data

I have a dataset of 1000 images. Every image is converted into YCbCr from RGB. It is found that the probability distribution of Cb, Cr for natural images is Gaussian. Now for the images that are present in my dataset I want to find out shape and scale parameters using maximum likelihood estimation in matlab. Can anyone please help me regarding this. By using those shape and scale parameters I want to find out the marginal distribution function for Gaussian distribution. The formulas are as follows:

P(Cb(x,y);σ_b,θ_b ) = θ_b/(2σ_b τ(1/θ_b )) exp(-|Cb(x,y)/σ_b |^(θ_b ))

P(Cr(x,y);σ_r,θ_r ) = θ_r/(2σ_r τ(1/θ_r )) exp(-|Cr(x,y)/σ_r |^(θ_r ))

where sigma(.) is a scale parameter and theta(.) is shape parameter. I have tried using mle function in matlab, but by using those parameter values the probability values I am getting are not in the range of 0 to 1 but they are increasing even beyond 1 giving very high values, which is not expected. Can anyone suggest me some remedy?