Diagnosing and monitoring vitiligo traditionally depend on visual assessments and manual tracing of depigmented patches on skin by medical professionals. This process, while widely adopted, is labor-intensive and challenging, especially when it comes to tracking disease progression or assessing treatment efficacy over time. This paper introduces an unsupervised learning approach for the automated segmentation of vitiligo-affected areas from skin images, circumventing the necessity for dataset-based training. We leverage a multi-stage processing pipeline that includes background exclusion, feature extraction, and image-based clustering. Qualitative analysis demonstrates the method's effectiveness and adaptability, with visual evidence highlighting the segmentation's robustness. Additionally, quantitative evaluation on 150 manually annotated images yields an average Intersection over Union (IoU) score of 0.47 and a Dice coefficient (DSC) score of 0.62, highlighting the method’s performance. The research underscores the potential of image-based unsupervised learning in dermatological imaging, paving the way for accelerated and accessible treatment planning and monitoring in vitiligo care.