Grayscale Image Enhancement Using SMQT (Successive Mean Quantization Transform) Algorithm

This MATLAB implementation is developed based on a research paper presented at the 2005 ICIP conference and includes comparative result visualizations. The code employs the SMQT (Successive Mean Quantization Transform) algorithm for grayscale image enhancement, which operates through recursive binary partitioning of pixel intensities around mean values. Unlike conventional histogram equalization that distorts the original histogram shape, SMQT preserves the histogram waveform while enhancing image contrast. The algorithm's core functionality involves multiple transformation levels where at each stage, pixel values are split into two groups based on their relationship to the current mean value, followed by independent processing of each subgroup. This hierarchical approach automatically concentrates on improving image details by emphasizing underlying structures through non-linear quantization. Compared to traditional histogram equalization methods, SMQT represents a novel grayscale enhancement technique that produces superior results by maintaining tonal balance while locally enhancing contrast in different intensity regions. The implementation includes functions for configuring quantization levels and visualizing before/after comparisons with histogram analysis.