Machine Learning Clustering Techniques for Segmenting IFS-Generated Garment Patterns

In the field of garment pattern creation, this work presents a contemporary implementation of Affine Iterated Function Systems (IFS). We use both K-means clustering and self-organizing map (SOM) to segment the visually appealing and complex patterns produced by IFS into discrete styles. These fractal patterns offer a break from traditional design techniques and are used as creative design templates for modern garment creation. Our research provides a cross-border new way of thinking in pattern design and prepares designers with new ways of seeking for originality, as it is an intersection of mathematical formulation and creative production. Box counting yields the fractal dimension of the created garment pattern, which is 1.477. This value indicates a balanced complexity appropriate for garment design. The resulting garments, inspired by fractal geometry, not only change aesthetic standards but also highlight the potential for sustainable and innovative practices in fashion.