From Elevation Maps To Contour Lines: SVM and Decision Trees to Detect Violin Width Reduction

2026-04-02

AI Summary

This study addresses the automatic detection of body-width reduction in three-dimensional violin models by systematically comparing, for the first time, the performance of support vector machines (SVM) and decision trees under two distinct input representations: raw geometric elevation maps and engineered features derived from parameterized contour lines. Leveraging mesh data acquired via 3D photogrammetry, the authors extract features through elevation map processing and contour line fitting. Experimental results demonstrate that the parameterized contour-based feature engineering approach significantly outperforms direct use of elevation maps, achieving higher accuracy and robustness in identifying width reductions. These findings validate the efficacy of traditional machine learning methods for detecting structural deformations in musical instruments and offer a novel methodological pathway for cultural heritage analysis.

Abstract

We explore the automatic detection of violin width reduction using 3D photogrammetric meshes. We compare SVM and Decision Trees applied to a geometry-based raw representation built from elevation maps with a more targeted, feature-engineered approach relying on parametric contour lines fitting. Although elevation maps occasionally achieve strong results, their performance does not surpass that of the contour-based inputs.

Problem

• violin width reduction
• 3D photogrammetric meshes
• elevation maps
• contour lines
• automatic detection

Innovation

• 3D photogrammetry
• elevation maps
• contour lines
• feature engineering
• violin width reduction