Anthropometry: Palm Size Clustering Using the K-Means Method and Silhouette Coefficient

Cucut Hariz Pratomo (Universitas Ahmad Dahlan)
Anton Yudhana (Universitas Ahmad Dahlan)
Herman (Universitas Ahmad Dahlan)

Jurnal Computech & Bisnis (e-Journal)
Vol. 20 No. 1 (2026): Jurnal Computech & Bisnis (e-Journal)

This study uses a palm image dataset to be used as training data and test data and the software used is the Python Programming Language and the OpenCV library which is quite light in processing on Google Collaboratory. The image data used is a palm image dataset from Kaggle which is open access to the public, the image data is then subjected to feature extraction using the Canny Edge Detection method with Euclidean distance calculations to see the edges of the object lines in the training image to clearly obtain the hand object that will be classified to see anthropometrically whether the palm is small, medium, and large. Grouping of 3 classes of palm size uses the K-Means method combined with the Silhouette Score to evaluate the results obtained. From a successful experiment, the Silhouette results show a score of 0.61, which means good.