Image Decomposition Using 2D Wavelet Decomposition Function with Code Implementation

Before performing image decomposition, we need to convert the normalized image data format from uint8 (optimized for image display) to double format (better suited for numerical computations). This conversion can be achieved using MATLAB's im2double() function or simply through data type casting. Subsequently, we call 2D wavelet decomposition functions such as wavedec2() in MATLAB, which implements multi-level wavelet decomposition using algorithms like Discrete Wavelet Transform (DWT) to separate image components across different frequency bands. To clearly visualize the pyramid structure of the decomposed image, we draw dividing lines in corresponding regions using functions like line() or plot(), enhancing both readability and visual representation of the decomposition results. The decomposition typically generates approximation coefficients (low-frequency components) and detail coefficients (horizontal, vertical, and diagonal high-frequency components) organized in a hierarchical structure.