Two-Dimensional Empirical Mode Decomposition (2D-EMD) Implementation

The current documentation provides technical specifications for the 2D Empirical Mode Decomposition implementation centered around Example.m. This primary script functions as the main controller that initiates the decomposition pipeline by invoking subsidiary functions and methods. The implementation follows the sifting process algorithm characteristic of EMD, which iteratively extracts oscillatory components from the input signal. The program requires 9.png as input data, which undergoes multidimensional signal processing to decompose into distinct modal components. When executed, Example.m automatically performs the Hilbert-Huang transform-based decomposition, separating the input image data into hierarchically organized intrinsic mode functions (IMFs) representing different frequency bands. The core algorithm implements envelope detection using spline interpolation and employs stopping criteria to ensure meaningful mode extraction. This decomposition methodology enables multiscale signal analysis through adaptive basis functions, particularly valuable for non-stationary image processing applications. The implementation handles boundary effects through mirror extension and incorporates mode splitting validation to prevent over-decomposition. Note that this represents an ongoing development project with regular updates to optimization algorithms and boundary handling techniques. Users should maintain current versions to leverage improved sifting stability and computational efficiency enhancements in the EMD framework.