Algorithm 3 MEMD-based image fusion algorithm.
1:	Convert all input images into a row (or column) vector; stack these vectors to form a multivariate signal, and apply MEMD;
2:	Reshape the resulting IMFs into its corresponding 2D images, yielding M sub-images ${\left\{I_m^p\right\}}_{m=1}^M$, for each input image;
3:	Compute the local activity level $A_m^p\left(a,b\right)$ for each sub-image $I_m^p\left(a,b\right)$ using Equations (4) or (5);
4:	Calculate local weighting factors $W{F}_m^p\left(a,b\right)$, based on $A_m^p\left(a,b\right)$, for each $I_m^p\left(a,b\right)$ using Equation (6);
5:	Obtain the m-th fused image Îm using Equation (7);
6:	Add all resulting fused images to obtain the final image using Equation (8).