Two-Dimensional Empirical Mode Decomposition

Two-Dimensional Empirical Mode Decomposition (EMD) is a signal decomposition method that decomposes signals into multiple Intrinsic Mode Functions (IMFs). The implementation employs morphological operations for extrema point detection, where structural elements are used to identify local maxima and minima across the signal surface. Subsequently, a radial basis function (RBF) neural network performs surface interpolation to fit the extracted extrema points, constructing the upper and lower envelopes through Gaussian kernel-based approximation. This approach achieves faster processing speeds by optimizing the envelope fitting procedure compared to traditional spline interpolation methods. The algorithm's efficiency makes it particularly suitable for decomposing complex 2D signals like images or spatial data patterns. Through two-dimensional EMD analysis, researchers can effectively extract and examine characteristic features and underlying properties of multidimensional signals.