Interactive Tool for a Multilevel Approach to Plagiarism Detection in Parametric 3D Mechanical CAD Models Based on Procedural Modeling Errors

Plagiarism detection in history-based parametric 3D CAD models remains challenging, particularly when models are altered or manually reconstructed. This study proposes a multilevel protocol and an interactive software tool to support the analysis of improper reuse through three complementary levels of evidence: administrative metadata evidence, geometric evidence, and procedural evidence derived from the modeling history, including potential error footprints based on the joint recurrence of multiple modeling errors. The protocol formalizes the relevant clues in a structured checklist and assigns weights to the indicators according to their relevance. The standalone application extracts and records the required metrics, performs pairwise comparisons, calculates administrative, geometric, and procedural figures of merit, and combines them into an overall similarity score. The approach provides quantitative and traceable evidence to support expert assessment of potential improper reuse. Results of our pilot experiments suggest that both the protocol and the application are viable and promising, although the weights may need additional refinement.