A Mean Curvature Approach to Boundary Detection: Geometric Insights for Unsupervised Learning

Alexandre Luis Magalhães Levada · 2026

Read A Mean Curvature Approach to Boundary Detection: Geometric Insights for Unsupervised Learning by Alexandre Luis Magalhães Levada (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-04274.txt.

Description

Alexandre L. M. Levada proposes a geometric way to find boundaries in high-dimensional data, a core task when grouping data without labels. Density-based methods can struggle when clusters are curved or unevenly spread, so he turns to the shape of the data cloud itself.

His method estimates local curvature from small neighbourhoods of points and treats high-curvature spots as boundaries, outliers, or transitions between groups. A percentile-based threshold then extracts boundaries at several scales without hand-set density parameters. The approach gives one geometric reading of features that are usually handled by separate tools.

Catalog record

CREATORAlexandre Luis Magalhães Levada
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE109 KB
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.04274

Browse the full library · About the Tower of Babel