Source Code for Calculating Fractal Box Dimension of Block-Based Grayscale Images

This source code implements the technique for calculating fractal box dimension of grayscale images through block partitioning. The implementation involves three main stages: first, the image is divided into non-overlapping blocks using a grid-based partitioning algorithm. Each block's grayscale values are then extracted and normalized for consistency. Next, the fractal box dimension algorithm is applied to each block using a counting method where boxes of varying sizes scan the grayscale data to determine dimension scaling properties. The core algorithm implements the relationship between box counts and scale sizes through logarithmic regression to compute dimension values. Finally, the individual block dimensions are aggregated using weighted averaging based on block significance, producing a comprehensive fractal dimension for the entire image. This technique provides fundamental quantitative analysis of grayscale image textures, serving as a basis for advanced image processing tasks like pattern recognition and texture classification. The code structure includes modular functions for image partitioning, grayscale processing, dimension calculation, and result integration.