Sasirangan fabric is a typical fabric from the South Kalimantan area. Sasirangan fabric patterns or motifs have a unique archetype that is different from other typical fabrics in Indonesia. The design of Sasirangan fabric is formed from the process of juju or seam. The pattern of Sasirangan fabric that has this uniqueness can be segmented into a more meaningful shape so that it is easy to analyze. The image segmentation that will be tested is the basic pattern of Sasirangan fabric with a random sample to compare the results of the evaluation of the metric evaluation of the image segmentation process from the Sasirangan fabric pattern. Image segmentation is a different segmentation with certain characteristics, namely using the compact watershed approach, canny filter, and morphological geodesic active contours method in the evaluation of image segmentation metrics using precision-recall, which serves to evaluate the quality of the classifier's output. After the image segmentation process is evaluated, the Sasirangan fabric pattern is grouped using the K-means algorithm as a different labelling strategy. This labelling process uses the K-means algorithm to better match details but can be unstable because it relies on random initialization. Alternatives to balance the unstable labelling process using the means algorithm can use discretization. The addition of the K-means method with discretization can create fields with geometric shapes that are pretty flat. The segmentation with Sasirangan fabric with a full motif or data number four 741.78s, results in processing the fastest and the longest computational time on data number two 120.79s.