Algorithm 3 Probabilistic PBOD
Require:	patch matching
	(a) Calculate 1 and 3 for all patches
	(b) Find βmax for both 1 and 3
	(c) Apply Neyman–Pearson to decide between RGB and HSV colour space
	If H1 is true
	${\beta }_6=\underset{\forall \beta }{\text{argmax}}{𝒟}_{P_3}$
	Else
	${\beta }_6=\underset{\forall \beta }{\text{argmax}}{𝒟}_{P_1}$
	End If
Ensure:	position smoothing
	(a) Select colour space based on εp
	(b) Determine the step size, δ
	(c) Construct the translated patch, $({\beta }_7^a,{\beta }_7^b,{\beta }_7^c,{\beta }_7^d)$
	(d) For each patch, histogram matching is modelled by Poisson distribution
	(e) Apply maximum likelihood for position adjustment
	(f) While β10 = β6 Do
	${\beta }_{10}=\underset{\forall {\beta }_9^i}{\text{argmax}}P(\mathbf{\text{x}}|{\beta }_9)$
	End While
Ensure:	size smoothing
	(a) Obtain prior probability by using Neyman–Pearson test
	(b) Obtain null hypothesis, H0
	$\begin{array}{cc}{max}_{\forall {\beta }^i}{P}_5(\mathbf{\text{x}}|{\beta }_i),& {\beta }_i\in \{{\beta }_{12}^a,\dots ,{\beta }_{12}^d\}\end{array}$
	(c) Obtain alternative hypothesis, H1
	$\begin{array}{cc}{max}_{\forall {\beta }^j}{P}_5(\mathbf{\text{x}}|{\beta }_j),& {\beta }_j\in \{{\beta }_{11}^a,\dots ,{\beta }_{11}^d\}\end{array}$
	(d) If H1 is favoured
	select 2nd set of priors
	Else
	select 1st set of priors
	End If
	(e) For each patch, obtain posterior probability by using Bayes risk
	(f) Normalize all histograms size for fairer comparison
	(g) If P5(β10|x) ≤ P5(βi|x), ${\beta }_i\in \{{\beta }_{11}^a,\dots ,{\beta }_{11}^d,{\beta }_{12}^a,\dots ,{\beta }_{12}^d\}$
	Size remain constant
	Else
	size is updated based on selected side
	End If    (h) Reiterate the process until the patch has converged to a certain size or number of iteration
	has exceeded p cycles.